diff --git a/docs/modelarts/umn/docker-modelarts_0029.html b/docs/modelarts/umn/docker-modelarts_0029.html
index a0535227..f9309c05 100644
--- a/docs/modelarts/umn/docker-modelarts_0029.html
+++ b/docs/modelarts/umn/docker-modelarts_0029.html
@@ -1,10 +1,10 @@
-
To migrate an image to the new training management version, perform the following operations:
-
- Add the default user group ma-group (GID = 100) for the image of the new-version training management. If the user group whose GID is 100 exists, skip this step.
- Add the default user ma-user (UID = 1000) for the image of the new-version training management. If the user whose UID is 1000 exists, skip this step.
- Modify the permission on files in the image to allow ma-user whose UID is 1000 to read and write the files.
-
You can modify an image by referring to the following Dockerfile so that the image complies with specifications for custom images of the new-version training management.
-
FROM {An existing image}
+To migrate an image to the new training management version, perform the following operations:
+
- Add the default user group ma-group (GID = 100) for the image of the new-version training management. If the user group whose GID is 100 exists, skip this step.
- Add the default user ma-user (UID = 1000) for the image of the new-version training management. If the user whose UID is 1000 exists, skip this step.
- Modify the permission on files in the image to allow ma-user whose UID is 1000 to read and write the files.
+
You can modify an image by referring to the following Dockerfile so that the image complies with specifications for custom images of the new-version training management.
+
FROM {An existing image}
USER root
@@ -16,14 +16,14 @@ ENV PYTHONUNBUFFERED=1
USER ma-user
WORKDIR /home/ma-user
-
In the new-version training management, custom images are executed by the user whose UID is 1000 by default. Before using an image in the new-version training management, check image file permissions on the server where the image was built. The process is as follows:
-
Run the following command to specify the user whose UID is 1000 to run the image:
-
docker run -u 1000 -ti {Custom image} bash
-
Run the boot command (such as the following command) in the container during image runtime:
-
python train.py
-
Check whether there are any Permission denied error.
-
If Permission denied is displayed for a file, for example, train.py, run the following command to change the file that belongs to the user whose UID is 1000:
-
chown 1000 train.py
+
In the new-version training management, custom images are executed by the user whose UID is 1000 by default. Before using an image in the new-version training management, check image file permissions on the server where the image was built. The process is as follows:
+
Run the following command to specify the user whose UID is 1000 to run the image:
+
docker run -u 1000 -ti {Custom image} bash
+
Run the boot command (such as the following command) in the container during image runtime:
+
python train.py
+
Check whether there are any Permission denied error.
+
If Permission denied is displayed for a file, for example, train.py, run the following command to change the file that belongs to the user whose UID is 1000:
+
chown 1000 train.py
diff --git a/docs/modelarts/umn/en-us_image_0000001799336916.png b/docs/modelarts/umn/en-us_image_0000001799336916.png
deleted file mode 100644
index d1f90ba1..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001799336916.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001799338328.png b/docs/modelarts/umn/en-us_image_0000001799338328.png
deleted file mode 100644
index de6bca1b..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001799338328.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001799338340.png b/docs/modelarts/umn/en-us_image_0000001799338340.png
deleted file mode 100644
index 8eea1bd2..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001799338340.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001799338500.png b/docs/modelarts/umn/en-us_image_0000001799338500.png
deleted file mode 100644
index 2ebdb21e..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001799338500.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001799338584.png b/docs/modelarts/umn/en-us_image_0000001799338584.png
deleted file mode 100644
index 9cf9fac2..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001799338584.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001799338620.png b/docs/modelarts/umn/en-us_image_0000001799338620.png
deleted file mode 100644
index ab29c560..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001799338620.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001799338648.png b/docs/modelarts/umn/en-us_image_0000001799338648.png
deleted file mode 100644
index 6859c74d..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001799338648.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001799338736.png b/docs/modelarts/umn/en-us_image_0000001799338736.png
deleted file mode 100644
index b668b1f9..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001799338736.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001799338788.png b/docs/modelarts/umn/en-us_image_0000001799338788.png
deleted file mode 100644
index 637e7a9f..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001799338788.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001799338864.png b/docs/modelarts/umn/en-us_image_0000001799338864.png
deleted file mode 100644
index ab890557..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001799338864.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001799339004.png b/docs/modelarts/umn/en-us_image_0000001799339004.png
deleted file mode 100644
index 748f1cca..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001799339004.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001799496664.png b/docs/modelarts/umn/en-us_image_0000001799496664.png
deleted file mode 100644
index 189e484b..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001799496664.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001799498100.png b/docs/modelarts/umn/en-us_image_0000001799498100.png
deleted file mode 100644
index 08cfb809..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001799498100.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001799498220.png b/docs/modelarts/umn/en-us_image_0000001799498220.png
deleted file mode 100644
index 4969ad2b..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001799498220.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001799498344.png b/docs/modelarts/umn/en-us_image_0000001799498344.png
deleted file mode 100644
index b642c55b..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001799498344.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001799498376.png b/docs/modelarts/umn/en-us_image_0000001799498376.png
deleted file mode 100644
index 45e10080..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001799498376.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001799498396.png b/docs/modelarts/umn/en-us_image_0000001799498396.png
deleted file mode 100644
index 9f9bd72f..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001799498396.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001799498592.png b/docs/modelarts/umn/en-us_image_0000001799498592.png
deleted file mode 100644
index ab1271d8..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001799498592.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001799498600.png b/docs/modelarts/umn/en-us_image_0000001799498600.png
deleted file mode 100644
index 3226dda5..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001799498600.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001799498620.png b/docs/modelarts/umn/en-us_image_0000001799498620.png
deleted file mode 100644
index bbdc1035..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001799498620.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001799498636.png b/docs/modelarts/umn/en-us_image_0000001799498636.png
deleted file mode 100644
index 8830411f..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001799498636.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001799498772.png b/docs/modelarts/umn/en-us_image_0000001799498772.png
deleted file mode 100644
index 97fd85e7..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001799498772.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001799498944.png b/docs/modelarts/umn/en-us_image_0000001799498944.png
deleted file mode 100644
index 5178669c..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001799498944.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001805228246.png b/docs/modelarts/umn/en-us_image_0000001805228246.png
deleted file mode 100644
index 1afe2c02..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001805228246.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001805277944.png b/docs/modelarts/umn/en-us_image_0000001805277944.png
deleted file mode 100644
index 77ff64a8..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001805277944.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001805279444.png b/docs/modelarts/umn/en-us_image_0000001805279444.png
deleted file mode 100644
index df10d119..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001805279444.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001805293966.png b/docs/modelarts/umn/en-us_image_0000001805293966.png
deleted file mode 100644
index 9d1e3572..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001805293966.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001805379758.png b/docs/modelarts/umn/en-us_image_0000001805379758.png
deleted file mode 100644
index f8ab16fb..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001805379758.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001805390986.png b/docs/modelarts/umn/en-us_image_0000001805390986.png
deleted file mode 100644
index c78557b5..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001805390986.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001805392998.png b/docs/modelarts/umn/en-us_image_0000001805392998.png
deleted file mode 100644
index 59e56102..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001805392998.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001805560062.png b/docs/modelarts/umn/en-us_image_0000001805560062.png
deleted file mode 100644
index 2596dee5..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001805560062.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001805560310.png b/docs/modelarts/umn/en-us_image_0000001805560310.png
deleted file mode 100644
index 9a4390f2..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001805560310.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001805745252.png b/docs/modelarts/umn/en-us_image_0000001805745252.png
deleted file mode 100644
index 1f4efabc..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001805745252.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001805914516.png b/docs/modelarts/umn/en-us_image_0000001805914516.png
deleted file mode 100644
index 21bbdd55..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001805914516.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001805917596.png b/docs/modelarts/umn/en-us_image_0000001805917596.png
deleted file mode 100644
index 21bbdd55..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001805917596.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001806111326.png b/docs/modelarts/umn/en-us_image_0000001806111326.png
deleted file mode 100644
index 4f7f446e..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001806111326.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001846057121.png b/docs/modelarts/umn/en-us_image_0000001846057121.png
deleted file mode 100644
index a00fec2d..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001846057121.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001846057173.png b/docs/modelarts/umn/en-us_image_0000001846057173.png
deleted file mode 100644
index d712a4aa..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001846057173.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001846057341.png b/docs/modelarts/umn/en-us_image_0000001846057341.png
deleted file mode 100644
index 8f892b0b..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001846057341.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001846057413.png b/docs/modelarts/umn/en-us_image_0000001846057413.png
deleted file mode 100644
index e5fd69a9..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001846057413.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001846057445.png b/docs/modelarts/umn/en-us_image_0000001846057445.png
deleted file mode 100644
index 9b9ccdaa..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001846057445.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001846057457.png b/docs/modelarts/umn/en-us_image_0000001846057457.png
deleted file mode 100644
index 206e4845..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001846057457.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001846057489.png b/docs/modelarts/umn/en-us_image_0000001846057489.png
deleted file mode 100644
index bdd161a0..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001846057489.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001846057505.png b/docs/modelarts/umn/en-us_image_0000001846057505.png
deleted file mode 100644
index 3a0168dd..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001846057505.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001846057601.png b/docs/modelarts/umn/en-us_image_0000001846057601.png
deleted file mode 100644
index b23e1126..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001846057601.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001846057725.png b/docs/modelarts/umn/en-us_image_0000001846057725.png
deleted file mode 100644
index 2b8a60d0..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001846057725.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001846058053.png b/docs/modelarts/umn/en-us_image_0000001846058053.png
deleted file mode 100644
index bb02e2d3..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001846058053.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001846137201.png b/docs/modelarts/umn/en-us_image_0000001846137201.png
deleted file mode 100644
index d49280fb..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001846137201.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001846137273.png b/docs/modelarts/umn/en-us_image_0000001846137273.png
deleted file mode 100644
index ff3a6b24..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001846137273.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001846137317.png b/docs/modelarts/umn/en-us_image_0000001846137317.png
deleted file mode 100644
index 953bf37c..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001846137317.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001846137593.png b/docs/modelarts/umn/en-us_image_0000001846137593.png
deleted file mode 100644
index 6436d8b7..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001846137593.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001846137605.png b/docs/modelarts/umn/en-us_image_0000001846137605.png
deleted file mode 100644
index fe583457..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001846137605.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001846137701.png b/docs/modelarts/umn/en-us_image_0000001846137701.png
deleted file mode 100644
index 953bf37c..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001846137701.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001846137705.png b/docs/modelarts/umn/en-us_image_0000001846137705.png
deleted file mode 100644
index 531431db..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001846137705.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001846137953.png b/docs/modelarts/umn/en-us_image_0000001846137953.png
deleted file mode 100644
index 4537818a..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001846137953.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001846138125.png b/docs/modelarts/umn/en-us_image_0000001846138125.png
deleted file mode 100644
index 4499c28a..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001846138125.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001851786845.png b/docs/modelarts/umn/en-us_image_0000001851786845.png
deleted file mode 100644
index f7b8eb69..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001851786845.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001851792033.png b/docs/modelarts/umn/en-us_image_0000001851792033.png
deleted file mode 100644
index e44ac1df..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001851792033.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001851841153.png b/docs/modelarts/umn/en-us_image_0000001851841153.png
deleted file mode 100644
index cfd4d7a6..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001851841153.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001851842725.png b/docs/modelarts/umn/en-us_image_0000001851842725.png
deleted file mode 100644
index 2b64193c..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001851842725.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001851867269.png b/docs/modelarts/umn/en-us_image_0000001851867269.png
deleted file mode 100644
index 7f5928da..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001851867269.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001851867781.png b/docs/modelarts/umn/en-us_image_0000001851867781.png
deleted file mode 100644
index 5b2e98d9..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001851867781.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001851916949.png b/docs/modelarts/umn/en-us_image_0000001851916949.png
deleted file mode 100644
index dd5db533..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001851916949.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001851919909.png b/docs/modelarts/umn/en-us_image_0000001851919909.png
deleted file mode 100644
index 919f37ce..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001851919909.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001851922269.png b/docs/modelarts/umn/en-us_image_0000001851922269.png
deleted file mode 100644
index ee7bdf06..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001851922269.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001852079633.png b/docs/modelarts/umn/en-us_image_0000001852079633.png
deleted file mode 100644
index c1284d41..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001852079633.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001852085129.png b/docs/modelarts/umn/en-us_image_0000001852085129.png
deleted file mode 100644
index e710210b..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001852085129.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001852149949.png b/docs/modelarts/umn/en-us_image_0000001852149949.png
deleted file mode 100644
index 4c5b3a04..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001852149949.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001852164713.png b/docs/modelarts/umn/en-us_image_0000001852164713.png
deleted file mode 100644
index c409e78b..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001852164713.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001852501737.png b/docs/modelarts/umn/en-us_image_0000001852501737.png
deleted file mode 100644
index c14af5d0..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001852501737.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001852592981.png b/docs/modelarts/umn/en-us_image_0000001852592981.png
deleted file mode 100644
index 5f6e9608..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001852592981.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001852636501.png b/docs/modelarts/umn/en-us_image_0000001852636501.png
deleted file mode 100644
index 21bbdd55..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001852636501.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001852712909.png b/docs/modelarts/umn/en-us_image_0000001852712909.png
deleted file mode 100644
index 232b6039..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001852712909.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001862605765.png b/docs/modelarts/umn/en-us_image_0000001862605765.png
deleted file mode 100644
index 29ff725d..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001862605765.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001862606185.png b/docs/modelarts/umn/en-us_image_0000001862606185.png
deleted file mode 100644
index 16183ae0..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001862606185.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001862682365.png b/docs/modelarts/umn/en-us_image_0000001862682365.png
deleted file mode 100644
index 7ca9c9a0..00000000
Binary files a/docs/modelarts/umn/en-us_image_0000001862682365.png and /dev/null differ
diff --git a/docs/modelarts/umn/en-us_image_0000001799337812.jpg b/docs/modelarts/umn/figure/en-us_image_0000001799337812.jpg
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799337812.jpg
rename to docs/modelarts/umn/figure/en-us_image_0000001799337812.jpg
diff --git a/docs/modelarts/umn/en-us_image_0000001799337980.png b/docs/modelarts/umn/figure/en-us_image_0000001799337980.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799337980.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799337980.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799338032.png b/docs/modelarts/umn/figure/en-us_image_0000001799338032.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799338032.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799338032.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799338144.png b/docs/modelarts/umn/figure/en-us_image_0000001799338144.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799338144.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799338144.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799338220.png b/docs/modelarts/umn/figure/en-us_image_0000001799338220.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799338220.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799338220.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799338228.png b/docs/modelarts/umn/figure/en-us_image_0000001799338228.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799338228.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799338228.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799338248.png b/docs/modelarts/umn/figure/en-us_image_0000001799338248.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799338248.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799338248.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799338300.png b/docs/modelarts/umn/figure/en-us_image_0000001799338300.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799338300.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799338300.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799338308.png b/docs/modelarts/umn/figure/en-us_image_0000001799338308.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799338308.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799338308.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799338312.png b/docs/modelarts/umn/figure/en-us_image_0000001799338312.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799338312.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799338312.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799338336.png b/docs/modelarts/umn/figure/en-us_image_0000001799338336.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799338336.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799338336.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799338352.png b/docs/modelarts/umn/figure/en-us_image_0000001799338352.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799338352.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799338352.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799338408.png b/docs/modelarts/umn/figure/en-us_image_0000001799338408.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799338408.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799338408.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799338424.png b/docs/modelarts/umn/figure/en-us_image_0000001799338424.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799338424.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799338424.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799338448.png b/docs/modelarts/umn/figure/en-us_image_0000001799338448.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799338448.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799338448.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799338464.png b/docs/modelarts/umn/figure/en-us_image_0000001799338464.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799338464.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799338464.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799338472.png b/docs/modelarts/umn/figure/en-us_image_0000001799338472.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799338472.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799338472.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799338488.png b/docs/modelarts/umn/figure/en-us_image_0000001799338488.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799338488.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799338488.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799338504.png b/docs/modelarts/umn/figure/en-us_image_0000001799338504.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799338504.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799338504.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799338536.png b/docs/modelarts/umn/figure/en-us_image_0000001799338536.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799338536.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799338536.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799338544.png b/docs/modelarts/umn/figure/en-us_image_0000001799338544.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799338544.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799338544.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799338560.png b/docs/modelarts/umn/figure/en-us_image_0000001799338560.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799338560.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799338560.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799338728.png b/docs/modelarts/umn/figure/en-us_image_0000001799338728.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799338728.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799338728.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799338848.png b/docs/modelarts/umn/figure/en-us_image_0000001799338848.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799338848.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799338848.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799338976.png b/docs/modelarts/umn/figure/en-us_image_0000001799338976.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799338976.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799338976.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799339188.png b/docs/modelarts/umn/figure/en-us_image_0000001799339188.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799339188.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799339188.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799497612.png b/docs/modelarts/umn/figure/en-us_image_0000001799497612.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799497612.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799497612.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799497936.png b/docs/modelarts/umn/figure/en-us_image_0000001799497936.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799497936.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799497936.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799497956.png b/docs/modelarts/umn/figure/en-us_image_0000001799497956.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799497956.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799497956.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799497968.png b/docs/modelarts/umn/figure/en-us_image_0000001799497968.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799497968.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799497968.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799498028.png b/docs/modelarts/umn/figure/en-us_image_0000001799498028.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799498028.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799498028.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799498036.png b/docs/modelarts/umn/figure/en-us_image_0000001799498036.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799498036.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799498036.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799498076.png b/docs/modelarts/umn/figure/en-us_image_0000001799498076.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799498076.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799498076.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799498084.png b/docs/modelarts/umn/figure/en-us_image_0000001799498084.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799498084.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799498084.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799498092.png b/docs/modelarts/umn/figure/en-us_image_0000001799498092.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799498092.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799498092.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799498096.png b/docs/modelarts/umn/figure/en-us_image_0000001799498096.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799498096.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799498096.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799498116.png b/docs/modelarts/umn/figure/en-us_image_0000001799498116.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799498116.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799498116.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799498136.png b/docs/modelarts/umn/figure/en-us_image_0000001799498136.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799498136.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799498136.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799498152.png b/docs/modelarts/umn/figure/en-us_image_0000001799498152.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799498152.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799498152.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799498176.png b/docs/modelarts/umn/figure/en-us_image_0000001799498176.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799498176.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799498176.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799498216.png b/docs/modelarts/umn/figure/en-us_image_0000001799498216.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799498216.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799498216.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799498232.png b/docs/modelarts/umn/figure/en-us_image_0000001799498232.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799498232.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799498232.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799498236.png b/docs/modelarts/umn/figure/en-us_image_0000001799498236.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799498236.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799498236.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799498252.png b/docs/modelarts/umn/figure/en-us_image_0000001799498252.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799498252.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799498252.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799498272.png b/docs/modelarts/umn/figure/en-us_image_0000001799498272.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799498272.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799498272.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799498288.png b/docs/modelarts/umn/figure/en-us_image_0000001799498288.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799498288.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799498288.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799498476.png b/docs/modelarts/umn/figure/en-us_image_0000001799498476.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799498476.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799498476.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799498484.png b/docs/modelarts/umn/figure/en-us_image_0000001799498484.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799498484.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799498484.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799498508.png b/docs/modelarts/umn/figure/en-us_image_0000001799498508.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799498508.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799498508.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799498612.png b/docs/modelarts/umn/figure/en-us_image_0000001799498612.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799498612.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799498612.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799499008.png b/docs/modelarts/umn/figure/en-us_image_0000001799499008.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799499008.png
rename to docs/modelarts/umn/figure/en-us_image_0000001799499008.png
diff --git a/docs/modelarts/umn/en-us_image_0000001805081250.png b/docs/modelarts/umn/figure/en-us_image_0000001805081250.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001805081250.png
rename to docs/modelarts/umn/figure/en-us_image_0000001805081250.png
diff --git a/docs/modelarts/umn/en-us_image_0000001805091682.png b/docs/modelarts/umn/figure/en-us_image_0000001805091682.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001805091682.png
rename to docs/modelarts/umn/figure/en-us_image_0000001805091682.png
diff --git a/docs/modelarts/umn/en-us_image_0000001805093186.png b/docs/modelarts/umn/figure/en-us_image_0000001805093186.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001805093186.png
rename to docs/modelarts/umn/figure/en-us_image_0000001805093186.png
diff --git a/docs/modelarts/umn/en-us_image_0000001805253002.png b/docs/modelarts/umn/figure/en-us_image_0000001805253002.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001805253002.png
rename to docs/modelarts/umn/figure/en-us_image_0000001805253002.png
diff --git a/docs/modelarts/umn/en-us_image_0000001805269900.png b/docs/modelarts/umn/figure/en-us_image_0000001805269900.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001805269900.png
rename to docs/modelarts/umn/figure/en-us_image_0000001805269900.png
diff --git a/docs/modelarts/umn/en-us_image_0000001805726206.png b/docs/modelarts/umn/figure/en-us_image_0000001805726206.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001805726206.png
rename to docs/modelarts/umn/figure/en-us_image_0000001805726206.png
diff --git a/docs/modelarts/umn/en-us_image_0000001806156600.png b/docs/modelarts/umn/figure/en-us_image_0000001806156600.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001806156600.png
rename to docs/modelarts/umn/figure/en-us_image_0000001806156600.png
diff --git a/docs/modelarts/umn/en-us_image_0000001806157356.png b/docs/modelarts/umn/figure/en-us_image_0000001806157356.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001806157356.png
rename to docs/modelarts/umn/figure/en-us_image_0000001806157356.png
diff --git a/docs/modelarts/umn/en-us_image_0000001806159588.png b/docs/modelarts/umn/figure/en-us_image_0000001806159588.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001806159588.png
rename to docs/modelarts/umn/figure/en-us_image_0000001806159588.png
diff --git a/docs/modelarts/umn/en-us_image_0000001806316500.png b/docs/modelarts/umn/figure/en-us_image_0000001806316500.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001806316500.png
rename to docs/modelarts/umn/figure/en-us_image_0000001806316500.png
diff --git a/docs/modelarts/umn/en-us_image_0000001806319384.png b/docs/modelarts/umn/figure/en-us_image_0000001806319384.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001806319384.png
rename to docs/modelarts/umn/figure/en-us_image_0000001806319384.png
diff --git a/docs/modelarts/umn/en-us_image_0000001809489970.png b/docs/modelarts/umn/figure/en-us_image_0000001809489970.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001809489970.png
rename to docs/modelarts/umn/figure/en-us_image_0000001809489970.png
diff --git a/docs/modelarts/umn/en-us_image_0000001809617018.png b/docs/modelarts/umn/figure/en-us_image_0000001809617018.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001809617018.png
rename to docs/modelarts/umn/figure/en-us_image_0000001809617018.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846057009.png b/docs/modelarts/umn/figure/en-us_image_0000001846057009.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846057009.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846057009.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846057085.png b/docs/modelarts/umn/figure/en-us_image_0000001846057085.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846057085.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846057085.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846057097.png b/docs/modelarts/umn/figure/en-us_image_0000001846057097.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846057097.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846057097.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846057113.png b/docs/modelarts/umn/figure/en-us_image_0000001846057113.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846057113.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846057113.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846057117.png b/docs/modelarts/umn/figure/en-us_image_0000001846057117.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846057117.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846057117.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846057129.png b/docs/modelarts/umn/figure/en-us_image_0000001846057129.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846057129.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846057129.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846057133.png b/docs/modelarts/umn/figure/en-us_image_0000001846057133.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846057133.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846057133.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846057137.png b/docs/modelarts/umn/figure/en-us_image_0000001846057137.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846057137.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846057137.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846057153.png b/docs/modelarts/umn/figure/en-us_image_0000001846057153.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846057153.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846057153.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846057157.png b/docs/modelarts/umn/figure/en-us_image_0000001846057157.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846057157.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846057157.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846057169.png b/docs/modelarts/umn/figure/en-us_image_0000001846057169.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846057169.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846057169.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846057177.png b/docs/modelarts/umn/figure/en-us_image_0000001846057177.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846057177.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846057177.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846057181.png b/docs/modelarts/umn/figure/en-us_image_0000001846057181.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846057181.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846057181.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846057189.png b/docs/modelarts/umn/figure/en-us_image_0000001846057189.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846057189.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846057189.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846057209.png b/docs/modelarts/umn/figure/en-us_image_0000001846057209.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846057209.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846057209.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846057237.png b/docs/modelarts/umn/figure/en-us_image_0000001846057237.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846057237.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846057237.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846057261.png b/docs/modelarts/umn/figure/en-us_image_0000001846057261.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846057261.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846057261.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846057317.png b/docs/modelarts/umn/figure/en-us_image_0000001846057317.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846057317.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846057317.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846057329.png b/docs/modelarts/umn/figure/en-us_image_0000001846057329.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846057329.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846057329.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846057337.png b/docs/modelarts/umn/figure/en-us_image_0000001846057337.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846057337.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846057337.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846057345.png b/docs/modelarts/umn/figure/en-us_image_0000001846057345.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846057345.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846057345.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846057353.png b/docs/modelarts/umn/figure/en-us_image_0000001846057353.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846057353.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846057353.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846057365.png b/docs/modelarts/umn/figure/en-us_image_0000001846057365.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846057365.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846057365.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846057393.png b/docs/modelarts/umn/figure/en-us_image_0000001846057393.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846057393.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846057393.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846057689.png b/docs/modelarts/umn/figure/en-us_image_0000001846057689.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846057689.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846057689.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846057741.png b/docs/modelarts/umn/figure/en-us_image_0000001846057741.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846057741.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846057741.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846058197.png b/docs/modelarts/umn/figure/en-us_image_0000001846058197.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846058197.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846058197.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846136833.png b/docs/modelarts/umn/figure/en-us_image_0000001846136833.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846136833.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846136833.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846136841.png b/docs/modelarts/umn/figure/en-us_image_0000001846136841.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846136841.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846136841.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846136977.png b/docs/modelarts/umn/figure/en-us_image_0000001846136977.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846136977.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846136977.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846137121.png b/docs/modelarts/umn/figure/en-us_image_0000001846137121.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846137121.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846137121.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846137129.png b/docs/modelarts/umn/figure/en-us_image_0000001846137129.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846137129.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846137129.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846137197.png b/docs/modelarts/umn/figure/en-us_image_0000001846137197.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846137197.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846137197.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846137217.png b/docs/modelarts/umn/figure/en-us_image_0000001846137217.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846137217.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846137217.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846137225.png b/docs/modelarts/umn/figure/en-us_image_0000001846137225.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846137225.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846137225.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846137269.png b/docs/modelarts/umn/figure/en-us_image_0000001846137269.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846137269.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846137269.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846137281.png b/docs/modelarts/umn/figure/en-us_image_0000001846137281.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846137281.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846137281.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846137289.png b/docs/modelarts/umn/figure/en-us_image_0000001846137289.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846137289.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846137289.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846137297.png b/docs/modelarts/umn/figure/en-us_image_0000001846137297.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846137297.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846137297.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846137325.png b/docs/modelarts/umn/figure/en-us_image_0000001846137325.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846137325.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846137325.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846137333.png b/docs/modelarts/umn/figure/en-us_image_0000001846137333.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846137333.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846137333.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846137341.png b/docs/modelarts/umn/figure/en-us_image_0000001846137341.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846137341.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846137341.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846137357.png b/docs/modelarts/umn/figure/en-us_image_0000001846137357.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846137357.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846137357.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846137373.png b/docs/modelarts/umn/figure/en-us_image_0000001846137373.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846137373.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846137373.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846137397.png b/docs/modelarts/umn/figure/en-us_image_0000001846137397.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846137397.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846137397.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846137413.png b/docs/modelarts/umn/figure/en-us_image_0000001846137413.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846137413.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846137413.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846137417.png b/docs/modelarts/umn/figure/en-us_image_0000001846137417.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846137417.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846137417.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846137433.png b/docs/modelarts/umn/figure/en-us_image_0000001846137433.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846137433.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846137433.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846137449.png b/docs/modelarts/umn/figure/en-us_image_0000001846137449.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846137449.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846137449.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846137469.png b/docs/modelarts/umn/figure/en-us_image_0000001846137469.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846137469.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846137469.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846137629.png b/docs/modelarts/umn/figure/en-us_image_0000001846137629.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846137629.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846137629.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846137769.png b/docs/modelarts/umn/figure/en-us_image_0000001846137769.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846137769.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846137769.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846137777.png b/docs/modelarts/umn/figure/en-us_image_0000001846137777.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846137777.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846137777.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846138101.png b/docs/modelarts/umn/figure/en-us_image_0000001846138101.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846138101.png
rename to docs/modelarts/umn/figure/en-us_image_0000001846138101.png
diff --git a/docs/modelarts/umn/en-us_image_0000001851800157.png b/docs/modelarts/umn/figure/en-us_image_0000001851800157.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001851800157.png
rename to docs/modelarts/umn/figure/en-us_image_0000001851800157.png
diff --git a/docs/modelarts/umn/en-us_image_0000001851873857.png b/docs/modelarts/umn/figure/en-us_image_0000001851873857.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001851873857.png
rename to docs/modelarts/umn/figure/en-us_image_0000001851873857.png
diff --git a/docs/modelarts/umn/en-us_image_0000001851879893.png b/docs/modelarts/umn/figure/en-us_image_0000001851879893.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001851879893.png
rename to docs/modelarts/umn/figure/en-us_image_0000001851879893.png
diff --git a/docs/modelarts/umn/en-us_image_0000001852878165.png b/docs/modelarts/umn/figure/en-us_image_0000001852878165.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001852878165.png
rename to docs/modelarts/umn/figure/en-us_image_0000001852878165.png
diff --git a/docs/modelarts/umn/en-us_image_0000001852879025.png b/docs/modelarts/umn/figure/en-us_image_0000001852879025.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001852879025.png
rename to docs/modelarts/umn/figure/en-us_image_0000001852879025.png
diff --git a/docs/modelarts/umn/en-us_image_0000001853022313.png b/docs/modelarts/umn/figure/en-us_image_0000001853022313.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001853022313.png
rename to docs/modelarts/umn/figure/en-us_image_0000001853022313.png
diff --git a/docs/modelarts/umn/en-us_image_0000001853035117.png b/docs/modelarts/umn/figure/en-us_image_0000001853035117.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001853035117.png
rename to docs/modelarts/umn/figure/en-us_image_0000001853035117.png
diff --git a/docs/modelarts/umn/en-us_image_0000001853037265.png b/docs/modelarts/umn/figure/en-us_image_0000001853037265.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001853037265.png
rename to docs/modelarts/umn/figure/en-us_image_0000001853037265.png
diff --git a/docs/modelarts/umn/en-us_image_0000001853038381.png b/docs/modelarts/umn/figure/en-us_image_0000001853038381.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001853038381.png
rename to docs/modelarts/umn/figure/en-us_image_0000001853038381.png
diff --git a/docs/modelarts/umn/en-us_image_0000001853038397.png b/docs/modelarts/umn/figure/en-us_image_0000001853038397.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001853038397.png
rename to docs/modelarts/umn/figure/en-us_image_0000001853038397.png
diff --git a/docs/modelarts/umn/en-us_image_0000001853040433.png b/docs/modelarts/umn/figure/en-us_image_0000001853040433.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001853040433.png
rename to docs/modelarts/umn/figure/en-us_image_0000001853040433.png
diff --git a/docs/modelarts/umn/en-us_image_0000001856209405.png b/docs/modelarts/umn/figure/en-us_image_0000001856209405.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001856209405.png
rename to docs/modelarts/umn/figure/en-us_image_0000001856209405.png
diff --git a/docs/modelarts/umn/en-us_image_0000001856210413.png b/docs/modelarts/umn/figure/en-us_image_0000001856210413.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001856210413.png
rename to docs/modelarts/umn/figure/en-us_image_0000001856210413.png
diff --git a/docs/modelarts/umn/en-us_image_0000001856316637.png b/docs/modelarts/umn/figure/en-us_image_0000001856316637.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001856316637.png
rename to docs/modelarts/umn/figure/en-us_image_0000001856316637.png
diff --git a/docs/modelarts/umn/en-us_image_0000001856322841.png b/docs/modelarts/umn/figure/en-us_image_0000001856322841.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001856322841.png
rename to docs/modelarts/umn/figure/en-us_image_0000001856322841.png
diff --git a/docs/modelarts/umn/en-us_image_0000001856327105.png b/docs/modelarts/umn/figure/en-us_image_0000001856327105.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001856327105.png
rename to docs/modelarts/umn/figure/en-us_image_0000001856327105.png
diff --git a/docs/modelarts/umn/en-us_image_0000001856406729.png b/docs/modelarts/umn/figure/en-us_image_0000001856406729.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001856406729.png
rename to docs/modelarts/umn/figure/en-us_image_0000001856406729.png
diff --git a/docs/modelarts/umn/en-us_image_0000001856410345.png b/docs/modelarts/umn/figure/en-us_image_0000001856410345.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001856410345.png
rename to docs/modelarts/umn/figure/en-us_image_0000001856410345.png
diff --git a/docs/modelarts/umn/en-us_image_0000001856413817.png b/docs/modelarts/umn/figure/en-us_image_0000001856413817.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001856413817.png
rename to docs/modelarts/umn/figure/en-us_image_0000001856413817.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846057481.png b/docs/modelarts/umn/figure/en-us_image_0000001909848844.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846057481.png
rename to docs/modelarts/umn/figure/en-us_image_0000001909848844.png
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001909848860.png b/docs/modelarts/umn/figure/en-us_image_0000001909848860.png
new file mode 100644
index 00000000..1d36ba92
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001909848860.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001909848864.png b/docs/modelarts/umn/figure/en-us_image_0000001909848864.png
new file mode 100644
index 00000000..e2da9525
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001909848864.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001909848892.png b/docs/modelarts/umn/figure/en-us_image_0000001909848892.png
new file mode 100644
index 00000000..b798ac1b
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001909848892.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001909848896.png b/docs/modelarts/umn/figure/en-us_image_0000001909848896.png
new file mode 100644
index 00000000..3c4cf0c3
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001909848896.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001909848908.png b/docs/modelarts/umn/figure/en-us_image_0000001909848908.png
new file mode 100644
index 00000000..70758d06
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001909848908.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001909848912.png b/docs/modelarts/umn/figure/en-us_image_0000001909848912.png
new file mode 100644
index 00000000..549550ef
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001909848912.png differ
diff --git a/docs/modelarts/umn/en-us_image_0000001846137661.png b/docs/modelarts/umn/figure/en-us_image_0000001909848924.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846137661.png
rename to docs/modelarts/umn/figure/en-us_image_0000001909848924.png
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001909849100.jpg b/docs/modelarts/umn/figure/en-us_image_0000001909849100.jpg
new file mode 100644
index 00000000..79c5ef27
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001909849100.jpg differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001909849108.png b/docs/modelarts/umn/figure/en-us_image_0000001909849108.png
new file mode 100644
index 00000000..a2b665db
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001909849108.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001909849112.png b/docs/modelarts/umn/figure/en-us_image_0000001909849112.png
new file mode 100644
index 00000000..e0cdcd70
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001909849112.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001909849128.png b/docs/modelarts/umn/figure/en-us_image_0000001909849128.png
new file mode 100644
index 00000000..694fbcf9
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001909849128.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001909849148.jpg b/docs/modelarts/umn/figure/en-us_image_0000001909849148.jpg
new file mode 100644
index 00000000..23911b65
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001909849148.jpg differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001909849152.jpg b/docs/modelarts/umn/figure/en-us_image_0000001909849152.jpg
new file mode 100644
index 00000000..faffea20
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001909849152.jpg differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001909849160.png b/docs/modelarts/umn/figure/en-us_image_0000001909849160.png
new file mode 100644
index 00000000..81271f91
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001909849160.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001909849180.jpg b/docs/modelarts/umn/figure/en-us_image_0000001909849180.jpg
new file mode 100644
index 00000000..d18e859f
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001909849180.jpg differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001909849228.png b/docs/modelarts/umn/figure/en-us_image_0000001909849228.png
new file mode 100644
index 00000000..eea73156
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001909849228.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001909849928.png b/docs/modelarts/umn/figure/en-us_image_0000001909849928.png
new file mode 100644
index 00000000..95399fe2
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001909849928.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910008856.png b/docs/modelarts/umn/figure/en-us_image_0000001910008856.png
new file mode 100644
index 00000000..93f4b6b2
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910008856.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910008860.png b/docs/modelarts/umn/figure/en-us_image_0000001910008860.png
new file mode 100644
index 00000000..2c3af174
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910008860.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910008868.png b/docs/modelarts/umn/figure/en-us_image_0000001910008868.png
new file mode 100644
index 00000000..649ff63d
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910008868.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910008900.png b/docs/modelarts/umn/figure/en-us_image_0000001910008900.png
new file mode 100644
index 00000000..cab670d5
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910008900.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910008908.png b/docs/modelarts/umn/figure/en-us_image_0000001910008908.png
new file mode 100644
index 00000000..927a4075
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910008908.png differ
diff --git a/docs/modelarts/umn/en-us_image_0000001799338768.png b/docs/modelarts/umn/figure/en-us_image_0000001910008912.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799338768.png
rename to docs/modelarts/umn/figure/en-us_image_0000001910008912.png
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910008920.png b/docs/modelarts/umn/figure/en-us_image_0000001910008920.png
new file mode 100644
index 00000000..7c0a8cfb
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910008920.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910008924.png b/docs/modelarts/umn/figure/en-us_image_0000001910008924.png
new file mode 100644
index 00000000..5df5094d
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910008924.png differ
diff --git a/docs/modelarts/umn/en-us_image_0000001846057389.png b/docs/modelarts/umn/figure/en-us_image_0000001910009104.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846057389.png
rename to docs/modelarts/umn/figure/en-us_image_0000001910009104.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799338988.png b/docs/modelarts/umn/figure/en-us_image_0000001910009112.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799338988.png
rename to docs/modelarts/umn/figure/en-us_image_0000001910009112.png
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910009120.png b/docs/modelarts/umn/figure/en-us_image_0000001910009120.png
new file mode 100644
index 00000000..ff60fc5b
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910009120.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910009156.jpg b/docs/modelarts/umn/figure/en-us_image_0000001910009156.jpg
new file mode 100644
index 00000000..dd352144
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910009156.jpg differ
diff --git a/docs/modelarts/umn/en-us_image_0000001799498016.png b/docs/modelarts/umn/figure/en-us_image_0000001910009168.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799498016.png
rename to docs/modelarts/umn/figure/en-us_image_0000001910009168.png
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910009188.jpg b/docs/modelarts/umn/figure/en-us_image_0000001910009188.jpg
new file mode 100644
index 00000000..d66463ce
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910009188.jpg differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910009236.png b/docs/modelarts/umn/figure/en-us_image_0000001910009236.png
new file mode 100644
index 00000000..74b1440b
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910009236.png differ
diff --git a/docs/modelarts/umn/en-us_image_0000001846058085.png b/docs/modelarts/umn/figure/en-us_image_0000001910009244.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846058085.png
rename to docs/modelarts/umn/figure/en-us_image_0000001910009244.png
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910009948.png b/docs/modelarts/umn/figure/en-us_image_0000001910009948.png
new file mode 100644
index 00000000..179b4a57
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910009948.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910009976.png b/docs/modelarts/umn/figure/en-us_image_0000001910009976.png
new file mode 100644
index 00000000..98d7ceeb
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910009976.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910016690.png b/docs/modelarts/umn/figure/en-us_image_0000001910016690.png
new file mode 100644
index 00000000..74e0e92a
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910016690.png differ
diff --git a/docs/modelarts/umn/en-us_image_0000001799339076.png b/docs/modelarts/umn/figure/en-us_image_0000001910016714.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799339076.png
rename to docs/modelarts/umn/figure/en-us_image_0000001910016714.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799498852.png b/docs/modelarts/umn/figure/en-us_image_0000001910016722.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799498852.png
rename to docs/modelarts/umn/figure/en-us_image_0000001910016722.png
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910018766.png b/docs/modelarts/umn/figure/en-us_image_0000001910018766.png
new file mode 100644
index 00000000..36cf2023
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910018766.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910018778.png b/docs/modelarts/umn/figure/en-us_image_0000001910018778.png
new file mode 100644
index 00000000..46206ae2
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910018778.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910018786.png b/docs/modelarts/umn/figure/en-us_image_0000001910018786.png
new file mode 100644
index 00000000..29434c14
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910018786.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910018790.png b/docs/modelarts/umn/figure/en-us_image_0000001910018790.png
new file mode 100644
index 00000000..a56505e9
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910018790.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910018798.png b/docs/modelarts/umn/figure/en-us_image_0000001910018798.png
new file mode 100644
index 00000000..7541b089
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910018798.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910018802.png b/docs/modelarts/umn/figure/en-us_image_0000001910018802.png
new file mode 100644
index 00000000..2004cfd4
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910018802.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910018922.png b/docs/modelarts/umn/figure/en-us_image_0000001910018922.png
new file mode 100644
index 00000000..0b4576a3
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910018922.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910018962.png b/docs/modelarts/umn/figure/en-us_image_0000001910018962.png
new file mode 100644
index 00000000..b22a13b7
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910018962.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910018966.png b/docs/modelarts/umn/figure/en-us_image_0000001910018966.png
new file mode 100644
index 00000000..ec75cfed
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910018966.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910018974.png b/docs/modelarts/umn/figure/en-us_image_0000001910018974.png
new file mode 100644
index 00000000..7a85b30d
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910018974.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910018978.png b/docs/modelarts/umn/figure/en-us_image_0000001910018978.png
new file mode 100644
index 00000000..49154fd3
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910018978.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910018986.png b/docs/modelarts/umn/figure/en-us_image_0000001910018986.png
new file mode 100644
index 00000000..c4efcc3c
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910018986.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910019018.png b/docs/modelarts/umn/figure/en-us_image_0000001910019018.png
new file mode 100644
index 00000000..3186a434
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910019018.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910019022.png b/docs/modelarts/umn/figure/en-us_image_0000001910019022.png
new file mode 100644
index 00000000..915f9566
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910019022.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910019046.png b/docs/modelarts/umn/figure/en-us_image_0000001910019046.png
new file mode 100644
index 00000000..7eee8739
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910019046.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910019074.png b/docs/modelarts/umn/figure/en-us_image_0000001910019074.png
new file mode 100644
index 00000000..3795979f
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910019074.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910019078.png b/docs/modelarts/umn/figure/en-us_image_0000001910019078.png
new file mode 100644
index 00000000..454f5ed6
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910019078.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910019094.png b/docs/modelarts/umn/figure/en-us_image_0000001910019094.png
new file mode 100644
index 00000000..f67ffbc7
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910019094.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910019098.png b/docs/modelarts/umn/figure/en-us_image_0000001910019098.png
new file mode 100644
index 00000000..7100a72f
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910019098.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910019102.png b/docs/modelarts/umn/figure/en-us_image_0000001910019102.png
new file mode 100644
index 00000000..eaf5d9e8
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910019102.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910019106.png b/docs/modelarts/umn/figure/en-us_image_0000001910019106.png
new file mode 100644
index 00000000..be0d9ebb
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910019106.png differ
diff --git a/docs/modelarts/umn/en-us_image_0000001799496692.png b/docs/modelarts/umn/figure/en-us_image_0000001910019742.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799496692.png
rename to docs/modelarts/umn/figure/en-us_image_0000001910019742.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799496656.png b/docs/modelarts/umn/figure/en-us_image_0000001910019746.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799496656.png
rename to docs/modelarts/umn/figure/en-us_image_0000001910019746.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799336884.png b/docs/modelarts/umn/figure/en-us_image_0000001910019770.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799336884.png
rename to docs/modelarts/umn/figure/en-us_image_0000001910019770.png
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910019930.png b/docs/modelarts/umn/figure/en-us_image_0000001910019930.png
new file mode 100644
index 00000000..bdba81bf
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910019930.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910023338.png b/docs/modelarts/umn/figure/en-us_image_0000001910023338.png
new file mode 100644
index 00000000..392a6fb3
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910023338.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910023346.png b/docs/modelarts/umn/figure/en-us_image_0000001910023346.png
new file mode 100644
index 00000000..79863ef8
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910023346.png differ
diff --git a/docs/modelarts/umn/en-us_image_0000001799498440.png b/docs/modelarts/umn/figure/en-us_image_0000001910027706.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799498440.png
rename to docs/modelarts/umn/figure/en-us_image_0000001910027706.png
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910027806.png b/docs/modelarts/umn/figure/en-us_image_0000001910027806.png
new file mode 100644
index 00000000..14aa22ca
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910027806.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910027814.png b/docs/modelarts/umn/figure/en-us_image_0000001910027814.png
new file mode 100644
index 00000000..373e2c41
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910027814.png differ
diff --git a/docs/modelarts/umn/en-us_image_0000001799498952.png b/docs/modelarts/umn/figure/en-us_image_0000001910027838.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799498952.png
rename to docs/modelarts/umn/figure/en-us_image_0000001910027838.png
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910027846.png b/docs/modelarts/umn/figure/en-us_image_0000001910027846.png
new file mode 100644
index 00000000..f09f9709
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910027846.png differ
diff --git a/docs/modelarts/umn/en-us_image_0000001846058025.png b/docs/modelarts/umn/figure/en-us_image_0000001910027850.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846058025.png
rename to docs/modelarts/umn/figure/en-us_image_0000001910027850.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799339208.png b/docs/modelarts/umn/figure/en-us_image_0000001910027858.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799339208.png
rename to docs/modelarts/umn/figure/en-us_image_0000001910027858.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846058037.png b/docs/modelarts/umn/figure/en-us_image_0000001910027866.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846058037.png
rename to docs/modelarts/umn/figure/en-us_image_0000001910027866.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846058045.png b/docs/modelarts/umn/figure/en-us_image_0000001910027874.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846058045.png
rename to docs/modelarts/umn/figure/en-us_image_0000001910027874.png
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910027890.png b/docs/modelarts/umn/figure/en-us_image_0000001910027890.png
new file mode 100644
index 00000000..0e7aac63
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910027890.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910027918.png b/docs/modelarts/umn/figure/en-us_image_0000001910027918.png
new file mode 100644
index 00000000..ac7d7566
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910027918.png differ
diff --git a/docs/modelarts/umn/en-us_image_0000001846138121.png b/docs/modelarts/umn/figure/en-us_image_0000001910027926.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846138121.png
rename to docs/modelarts/umn/figure/en-us_image_0000001910027926.png
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910027930.png b/docs/modelarts/umn/figure/en-us_image_0000001910027930.png
new file mode 100644
index 00000000..c1360a80
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910027930.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910027934.png b/docs/modelarts/umn/figure/en-us_image_0000001910027934.png
new file mode 100644
index 00000000..ad140a54
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910027934.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910027938.png b/docs/modelarts/umn/figure/en-us_image_0000001910027938.png
new file mode 100644
index 00000000..517298b5
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910027938.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910053574.png b/docs/modelarts/umn/figure/en-us_image_0000001910053574.png
new file mode 100644
index 00000000..0c2e4204
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910053574.png differ
diff --git a/docs/modelarts/umn/en-us_image_0000001799338368.png b/docs/modelarts/umn/figure/en-us_image_0000001910055010.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799338368.png
rename to docs/modelarts/umn/figure/en-us_image_0000001910055010.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846057865.png b/docs/modelarts/umn/figure/en-us_image_0000001910055046.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846057865.png
rename to docs/modelarts/umn/figure/en-us_image_0000001910055046.png
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910055054.png b/docs/modelarts/umn/figure/en-us_image_0000001910055054.png
new file mode 100644
index 00000000..a02518a0
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910055054.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910055454.png b/docs/modelarts/umn/figure/en-us_image_0000001910055454.png
new file mode 100644
index 00000000..bcc7559e
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910055454.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910056710.png b/docs/modelarts/umn/figure/en-us_image_0000001910056710.png
new file mode 100644
index 00000000..14efe874
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910056710.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910056718.png b/docs/modelarts/umn/figure/en-us_image_0000001910056718.png
new file mode 100644
index 00000000..a0754553
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910056718.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910056730.png b/docs/modelarts/umn/figure/en-us_image_0000001910056730.png
new file mode 100644
index 00000000..ab35fc34
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910056730.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910058794.png b/docs/modelarts/umn/figure/en-us_image_0000001910058794.png
new file mode 100644
index 00000000..135e076d
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910058794.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910058806.png b/docs/modelarts/umn/figure/en-us_image_0000001910058806.png
new file mode 100644
index 00000000..d2d52b76
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910058806.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910058810.png b/docs/modelarts/umn/figure/en-us_image_0000001910058810.png
new file mode 100644
index 00000000..f23377c4
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910058810.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910058818.png b/docs/modelarts/umn/figure/en-us_image_0000001910058818.png
new file mode 100644
index 00000000..6fc5ab17
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910058818.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910058950.png b/docs/modelarts/umn/figure/en-us_image_0000001910058950.png
new file mode 100644
index 00000000..4505d3c1
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910058950.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910058954.png b/docs/modelarts/umn/figure/en-us_image_0000001910058954.png
new file mode 100644
index 00000000..74d01268
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910058954.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910058978.png b/docs/modelarts/umn/figure/en-us_image_0000001910058978.png
new file mode 100644
index 00000000..65e65328
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910058978.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910058990.png b/docs/modelarts/umn/figure/en-us_image_0000001910058990.png
new file mode 100644
index 00000000..d36ba375
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910058990.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910059034.png b/docs/modelarts/umn/figure/en-us_image_0000001910059034.png
new file mode 100644
index 00000000..4c02329a
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910059034.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910059042.png b/docs/modelarts/umn/figure/en-us_image_0000001910059042.png
new file mode 100644
index 00000000..00bc6834
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910059042.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910059086.png b/docs/modelarts/umn/figure/en-us_image_0000001910059086.png
new file mode 100644
index 00000000..15fd9a09
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910059086.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910059090.png b/docs/modelarts/umn/figure/en-us_image_0000001910059090.png
new file mode 100644
index 00000000..377d1602
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910059090.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910059094.png b/docs/modelarts/umn/figure/en-us_image_0000001910059094.png
new file mode 100644
index 00000000..ab2068ab
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910059094.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910059106.png b/docs/modelarts/umn/figure/en-us_image_0000001910059106.png
new file mode 100644
index 00000000..0a724179
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910059106.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910059110.png b/docs/modelarts/umn/figure/en-us_image_0000001910059110.png
new file mode 100644
index 00000000..56cf33ac
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910059110.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910059114.png b/docs/modelarts/umn/figure/en-us_image_0000001910059114.png
new file mode 100644
index 00000000..f9f450ce
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910059114.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910059118.png b/docs/modelarts/umn/figure/en-us_image_0000001910059118.png
new file mode 100644
index 00000000..dd9b87b4
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910059118.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910059122.png b/docs/modelarts/umn/figure/en-us_image_0000001910059122.png
new file mode 100644
index 00000000..5a283318
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910059122.png differ
diff --git a/docs/modelarts/umn/en-us_image_0000001846055741.png b/docs/modelarts/umn/figure/en-us_image_0000001910059738.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846055741.png
rename to docs/modelarts/umn/figure/en-us_image_0000001910059738.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846135813.png b/docs/modelarts/umn/figure/en-us_image_0000001910059742.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846135813.png
rename to docs/modelarts/umn/figure/en-us_image_0000001910059742.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799336880.png b/docs/modelarts/umn/figure/en-us_image_0000001910059758.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799336880.png
rename to docs/modelarts/umn/figure/en-us_image_0000001910059758.png
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910059926.png b/docs/modelarts/umn/figure/en-us_image_0000001910059926.png
new file mode 100644
index 00000000..8e535f20
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910059926.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910063354.png b/docs/modelarts/umn/figure/en-us_image_0000001910063354.png
new file mode 100644
index 00000000..7d4cc5ea
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910063354.png differ
diff --git a/docs/modelarts/umn/en-us_image_0000001846057669.png b/docs/modelarts/umn/figure/en-us_image_0000001910067714.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846057669.png
rename to docs/modelarts/umn/figure/en-us_image_0000001910067714.png
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910067778.png b/docs/modelarts/umn/figure/en-us_image_0000001910067778.png
new file mode 100644
index 00000000..5d0b7143
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910067778.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910067790.png b/docs/modelarts/umn/figure/en-us_image_0000001910067790.png
new file mode 100644
index 00000000..38ddf54d
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910067790.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910067794.png b/docs/modelarts/umn/figure/en-us_image_0000001910067794.png
new file mode 100644
index 00000000..4c066dd1
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910067794.png differ
diff --git a/docs/modelarts/umn/en-us_image_0000001846138141.png b/docs/modelarts/umn/figure/en-us_image_0000001910067818.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846138141.png
rename to docs/modelarts/umn/figure/en-us_image_0000001910067818.png
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910067822.png b/docs/modelarts/umn/figure/en-us_image_0000001910067822.png
new file mode 100644
index 00000000..3af79984
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910067822.png differ
diff --git a/docs/modelarts/umn/en-us_image_0000001846138105.png b/docs/modelarts/umn/figure/en-us_image_0000001910067846.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846138105.png
rename to docs/modelarts/umn/figure/en-us_image_0000001910067846.png
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910067874.png b/docs/modelarts/umn/figure/en-us_image_0000001910067874.png
new file mode 100644
index 00000000..9e1c9f7a
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910067874.png differ
diff --git a/docs/modelarts/umn/en-us_image_0000001799339156.png b/docs/modelarts/umn/figure/en-us_image_0000001910067878.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799339156.png
rename to docs/modelarts/umn/figure/en-us_image_0000001910067878.png
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910067906.png b/docs/modelarts/umn/figure/en-us_image_0000001910067906.png
new file mode 100644
index 00000000..b155bec6
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910067906.png differ
diff --git a/docs/modelarts/umn/en-us_image_0000001799339184.png b/docs/modelarts/umn/figure/en-us_image_0000001910067918.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799339184.png
rename to docs/modelarts/umn/figure/en-us_image_0000001910067918.png
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910067926.png b/docs/modelarts/umn/figure/en-us_image_0000001910067926.png
new file mode 100644
index 00000000..bad8c159
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910067926.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001910067930.png b/docs/modelarts/umn/figure/en-us_image_0000001910067930.png
new file mode 100644
index 00000000..e160d4f6
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001910067930.png differ
diff --git a/docs/modelarts/umn/en-us_image_0000001799338580.png b/docs/modelarts/umn/figure/en-us_image_0000001910067970.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799338580.png
rename to docs/modelarts/umn/figure/en-us_image_0000001910067970.png
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001914601386.png b/docs/modelarts/umn/figure/en-us_image_0000001914601386.png
new file mode 100644
index 00000000..228dd454
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001914601386.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001914602074.png b/docs/modelarts/umn/figure/en-us_image_0000001914602074.png
new file mode 100644
index 00000000..bc8e34fa
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001914602074.png differ
diff --git a/docs/modelarts/umn/en-us_image_0000001851868941.png b/docs/modelarts/umn/figure/en-us_image_0000001914973738.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001851868941.png
rename to docs/modelarts/umn/figure/en-us_image_0000001914973738.png
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001916187730.png b/docs/modelarts/umn/figure/en-us_image_0000001916187730.png
new file mode 100644
index 00000000..228dd454
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001916187730.png differ
diff --git a/docs/modelarts/umn/en-us_image_0000001852057389.png b/docs/modelarts/umn/figure/en-us_image_0000001916222240.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001852057389.png
rename to docs/modelarts/umn/figure/en-us_image_0000001916222240.png
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001916347638.png b/docs/modelarts/umn/figure/en-us_image_0000001916347638.png
new file mode 100644
index 00000000..2f2a9624
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001916347638.png differ
diff --git a/docs/modelarts/umn/en-us_image_0000001852060437.png b/docs/modelarts/umn/figure/en-us_image_0000001916382140.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001852060437.png
rename to docs/modelarts/umn/figure/en-us_image_0000001916382140.png
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001916390764.png b/docs/modelarts/umn/figure/en-us_image_0000001916390764.png
new file mode 100644
index 00000000..0e8ed845
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001916390764.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001916395064.png b/docs/modelarts/umn/figure/en-us_image_0000001916395064.png
new file mode 100644
index 00000000..b0ce8fc4
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001916395064.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001916397372.png b/docs/modelarts/umn/figure/en-us_image_0000001916397372.png
new file mode 100644
index 00000000..d9cca129
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001916397372.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001916469920.png b/docs/modelarts/umn/figure/en-us_image_0000001916469920.png
new file mode 100644
index 00000000..05f8d33f
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001916469920.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001916629192.png b/docs/modelarts/umn/figure/en-us_image_0000001916629192.png
new file mode 100644
index 00000000..8aad52c1
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001916629192.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001916629860.png b/docs/modelarts/umn/figure/en-us_image_0000001916629860.png
new file mode 100644
index 00000000..45fc45be
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001916629860.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001916630068.png b/docs/modelarts/umn/figure/en-us_image_0000001916630068.png
new file mode 100644
index 00000000..8407e016
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001916630068.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001916953708.png b/docs/modelarts/umn/figure/en-us_image_0000001916953708.png
new file mode 100644
index 00000000..b093f848
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001916953708.png differ
diff --git a/docs/modelarts/umn/en-us_image_0000001799336924.png b/docs/modelarts/umn/figure/en-us_image_0000001919126620.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799336924.png
rename to docs/modelarts/umn/figure/en-us_image_0000001919126620.png
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001919403074.png b/docs/modelarts/umn/figure/en-us_image_0000001919403074.png
new file mode 100644
index 00000000..ac841993
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001919403074.png differ
diff --git a/docs/modelarts/umn/en-us_image_0000001851792857.png b/docs/modelarts/umn/figure/en-us_image_0000001919561582.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001851792857.png
rename to docs/modelarts/umn/figure/en-us_image_0000001919561582.png
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001919562630.png b/docs/modelarts/umn/figure/en-us_image_0000001919562630.png
new file mode 100644
index 00000000..595f7dbd
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001919562630.png differ
diff --git a/docs/modelarts/umn/en-us_image_0000001799339192.png b/docs/modelarts/umn/figure/en-us_image_0000001943968053.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799339192.png
rename to docs/modelarts/umn/figure/en-us_image_0000001943968053.png
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943968061.png b/docs/modelarts/umn/figure/en-us_image_0000001943968061.png
new file mode 100644
index 00000000..25558939
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943968061.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943968081.png b/docs/modelarts/umn/figure/en-us_image_0000001943968081.png
new file mode 100644
index 00000000..33b46ae2
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943968081.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943968105.png b/docs/modelarts/umn/figure/en-us_image_0000001943968105.png
new file mode 100644
index 00000000..a831581e
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943968105.png differ
diff --git a/docs/modelarts/umn/en-us_image_0000001799498528.png b/docs/modelarts/umn/figure/en-us_image_0000001943968109.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799498528.png
rename to docs/modelarts/umn/figure/en-us_image_0000001943968109.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846057625.png b/docs/modelarts/umn/figure/en-us_image_0000001943968113.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846057625.png
rename to docs/modelarts/umn/figure/en-us_image_0000001943968113.png
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943968121.png b/docs/modelarts/umn/figure/en-us_image_0000001943968121.png
new file mode 100644
index 00000000..5fdef87a
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943968121.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943968129.png b/docs/modelarts/umn/figure/en-us_image_0000001943968129.png
new file mode 100644
index 00000000..f0031f9e
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943968129.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943968301.jpg b/docs/modelarts/umn/figure/en-us_image_0000001943968301.jpg
new file mode 100644
index 00000000..4e993b53
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943968301.jpg differ
diff --git a/docs/modelarts/umn/en-us_image_0000001799498840.png b/docs/modelarts/umn/figure/en-us_image_0000001943968305.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799498840.png
rename to docs/modelarts/umn/figure/en-us_image_0000001943968305.png
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943968313.png b/docs/modelarts/umn/figure/en-us_image_0000001943968313.png
new file mode 100644
index 00000000..30f15fd6
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943968313.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943968337.png b/docs/modelarts/umn/figure/en-us_image_0000001943968337.png
new file mode 100644
index 00000000..18c26719
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943968337.png differ
diff --git a/docs/modelarts/umn/en-us_image_0000001846137453.png b/docs/modelarts/umn/figure/en-us_image_0000001943968341.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846137453.png
rename to docs/modelarts/umn/figure/en-us_image_0000001943968341.png
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943968345.jpg b/docs/modelarts/umn/figure/en-us_image_0000001943968345.jpg
new file mode 100644
index 00000000..77943442
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943968345.jpg differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943968361.png b/docs/modelarts/umn/figure/en-us_image_0000001943968361.png
new file mode 100644
index 00000000..2f1bd7a2
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943968361.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943968397.png b/docs/modelarts/umn/figure/en-us_image_0000001943968397.png
new file mode 100644
index 00000000..45dd2ac8
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943968397.png differ
diff --git a/docs/modelarts/umn/en-us_image_0000001846137653.png b/docs/modelarts/umn/figure/en-us_image_0000001943968425.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846137653.png
rename to docs/modelarts/umn/figure/en-us_image_0000001943968425.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799498996.png b/docs/modelarts/umn/figure/en-us_image_0000001943968433.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799498996.png
rename to docs/modelarts/umn/figure/en-us_image_0000001943968433.png
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943969161.png b/docs/modelarts/umn/figure/en-us_image_0000001943969161.png
new file mode 100644
index 00000000..279dbe4f
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943969161.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943969601.png b/docs/modelarts/umn/figure/en-us_image_0000001943969601.png
new file mode 100644
index 00000000..7c590bba
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943969601.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943972257.png b/docs/modelarts/umn/figure/en-us_image_0000001943972257.png
new file mode 100644
index 00000000..7c5c765a
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943972257.png differ
diff --git a/docs/modelarts/umn/en-us_image_0000001846138133.png b/docs/modelarts/umn/figure/en-us_image_0000001943972477.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846138133.png
rename to docs/modelarts/umn/figure/en-us_image_0000001943972477.png
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943972845.png b/docs/modelarts/umn/figure/en-us_image_0000001943972845.png
new file mode 100644
index 00000000..15412d16
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943972845.png differ
diff --git a/docs/modelarts/umn/en-us_image_0000001805757768.png b/docs/modelarts/umn/figure/en-us_image_0000001943974229.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001805757768.png
rename to docs/modelarts/umn/figure/en-us_image_0000001943974229.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799339008.png b/docs/modelarts/umn/figure/en-us_image_0000001943974253.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799339008.png
rename to docs/modelarts/umn/figure/en-us_image_0000001943974253.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799499024.png b/docs/modelarts/umn/figure/en-us_image_0000001943974333.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799499024.png
rename to docs/modelarts/umn/figure/en-us_image_0000001943974333.png
diff --git a/docs/modelarts/umn/en-us_image_0000001846057553.png b/docs/modelarts/umn/figure/en-us_image_0000001943974441.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846057553.png
rename to docs/modelarts/umn/figure/en-us_image_0000001943974441.png
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943975617.png b/docs/modelarts/umn/figure/en-us_image_0000001943975617.png
new file mode 100644
index 00000000..97b4da54
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943975617.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943975885.png b/docs/modelarts/umn/figure/en-us_image_0000001943975885.png
new file mode 100644
index 00000000..e1370b38
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943975885.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943975893.png b/docs/modelarts/umn/figure/en-us_image_0000001943975893.png
new file mode 100644
index 00000000..a25cde0d
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943975893.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943975897.png b/docs/modelarts/umn/figure/en-us_image_0000001943975897.png
new file mode 100644
index 00000000..9133344e
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943975897.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943977929.png b/docs/modelarts/umn/figure/en-us_image_0000001943977929.png
new file mode 100644
index 00000000..fb0f4db8
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943977929.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943977961.png b/docs/modelarts/umn/figure/en-us_image_0000001943977961.png
new file mode 100644
index 00000000..e9aad7ed
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943977961.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943977985.png b/docs/modelarts/umn/figure/en-us_image_0000001943977985.png
new file mode 100644
index 00000000..d456c6a9
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943977985.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943977993.png b/docs/modelarts/umn/figure/en-us_image_0000001943977993.png
new file mode 100644
index 00000000..f294cf68
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943977993.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943978033.png b/docs/modelarts/umn/figure/en-us_image_0000001943978033.png
new file mode 100644
index 00000000..74a47fed
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943978033.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943978049.png b/docs/modelarts/umn/figure/en-us_image_0000001943978049.png
new file mode 100644
index 00000000..2613647c
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943978049.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943978149.png b/docs/modelarts/umn/figure/en-us_image_0000001943978149.png
new file mode 100644
index 00000000..ba79c73c
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943978149.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943978153.png b/docs/modelarts/umn/figure/en-us_image_0000001943978153.png
new file mode 100644
index 00000000..ff0f4961
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943978153.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943978157.png b/docs/modelarts/umn/figure/en-us_image_0000001943978157.png
new file mode 100644
index 00000000..1c2bce29
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943978157.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943978209.png b/docs/modelarts/umn/figure/en-us_image_0000001943978209.png
new file mode 100644
index 00000000..2ef5021d
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943978209.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943978237.png b/docs/modelarts/umn/figure/en-us_image_0000001943978237.png
new file mode 100644
index 00000000..aa30b3bf
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943978237.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943978261.png b/docs/modelarts/umn/figure/en-us_image_0000001943978261.png
new file mode 100644
index 00000000..27792b0c
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943978261.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943978265.png b/docs/modelarts/umn/figure/en-us_image_0000001943978265.png
new file mode 100644
index 00000000..93660b2b
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943978265.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943978277.png b/docs/modelarts/umn/figure/en-us_image_0000001943978277.png
new file mode 100644
index 00000000..7533e10f
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943978277.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943978281.png b/docs/modelarts/umn/figure/en-us_image_0000001943978281.png
new file mode 100644
index 00000000..376919ec
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943978281.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943978285.png b/docs/modelarts/umn/figure/en-us_image_0000001943978285.png
new file mode 100644
index 00000000..42e8cc16
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943978285.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943978289.png b/docs/modelarts/umn/figure/en-us_image_0000001943978289.png
new file mode 100644
index 00000000..6fc5ab17
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943978289.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943978293.png b/docs/modelarts/umn/figure/en-us_image_0000001943978293.png
new file mode 100644
index 00000000..5fdef87a
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943978293.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943978297.png b/docs/modelarts/umn/figure/en-us_image_0000001943978297.png
new file mode 100644
index 00000000..0c2e4204
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943978297.png differ
diff --git a/docs/modelarts/umn/en-us_image_0000001846055693.png b/docs/modelarts/umn/figure/en-us_image_0000001943978949.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846055693.png
rename to docs/modelarts/umn/figure/en-us_image_0000001943978949.png
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943979113.png b/docs/modelarts/umn/figure/en-us_image_0000001943979113.png
new file mode 100644
index 00000000..b0b6403b
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943979113.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943979185.png b/docs/modelarts/umn/figure/en-us_image_0000001943979185.png
new file mode 100644
index 00000000..8df2101c
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943979185.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943982445.png b/docs/modelarts/umn/figure/en-us_image_0000001943982445.png
new file mode 100644
index 00000000..abfb8ec8
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943982445.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943982549.png b/docs/modelarts/umn/figure/en-us_image_0000001943982549.png
new file mode 100644
index 00000000..255a8eee
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943982549.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943986949.png b/docs/modelarts/umn/figure/en-us_image_0000001943986949.png
new file mode 100644
index 00000000..e9711b1d
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943986949.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943987005.png b/docs/modelarts/umn/figure/en-us_image_0000001943987005.png
new file mode 100644
index 00000000..9cad175c
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943987005.png differ
diff --git a/docs/modelarts/umn/en-us_image_0000001799498988.png b/docs/modelarts/umn/figure/en-us_image_0000001943987009.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799498988.png
rename to docs/modelarts/umn/figure/en-us_image_0000001943987009.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799498964.png b/docs/modelarts/umn/figure/en-us_image_0000001943987017.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799498964.png
rename to docs/modelarts/umn/figure/en-us_image_0000001943987017.png
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943987025.png b/docs/modelarts/umn/figure/en-us_image_0000001943987025.png
new file mode 100644
index 00000000..56c075f8
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943987025.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943987037.png b/docs/modelarts/umn/figure/en-us_image_0000001943987037.png
new file mode 100644
index 00000000..8708ea08
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943987037.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943987045.png b/docs/modelarts/umn/figure/en-us_image_0000001943987045.png
new file mode 100644
index 00000000..58bc31db
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943987045.png differ
diff --git a/docs/modelarts/umn/en-us_image_0000001799498948.png b/docs/modelarts/umn/figure/en-us_image_0000001943987053.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799498948.png
rename to docs/modelarts/umn/figure/en-us_image_0000001943987053.png
diff --git a/docs/modelarts/umn/en-us_image_0000001799339200.png b/docs/modelarts/umn/figure/en-us_image_0000001943987057.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799339200.png
rename to docs/modelarts/umn/figure/en-us_image_0000001943987057.png
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943987085.png b/docs/modelarts/umn/figure/en-us_image_0000001943987085.png
new file mode 100644
index 00000000..42441b2f
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943987085.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943987113.png b/docs/modelarts/umn/figure/en-us_image_0000001943987113.png
new file mode 100644
index 00000000..4f6d3dca
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943987113.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943987117.png b/docs/modelarts/umn/figure/en-us_image_0000001943987117.png
new file mode 100644
index 00000000..a3a91d40
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943987117.png differ
diff --git a/docs/modelarts/umn/en-us_image_0000001846058061.png b/docs/modelarts/umn/figure/en-us_image_0000001943987121.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001846058061.png
rename to docs/modelarts/umn/figure/en-us_image_0000001943987121.png
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001943987133.png b/docs/modelarts/umn/figure/en-us_image_0000001943987133.png
new file mode 100644
index 00000000..75490eea
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001943987133.png differ
diff --git a/docs/modelarts/umn/en-us_image_0000001799498072.png b/docs/modelarts/umn/figure/en-us_image_0000001943987149.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001799498072.png
rename to docs/modelarts/umn/figure/en-us_image_0000001943987149.png
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001946316357.png b/docs/modelarts/umn/figure/en-us_image_0000001946316357.png
new file mode 100644
index 00000000..2f2a9624
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001946316357.png differ
diff --git a/docs/modelarts/umn/en-us_image_0000001805229686.png b/docs/modelarts/umn/figure/en-us_image_0000001946372881.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001805229686.png
rename to docs/modelarts/umn/figure/en-us_image_0000001946372881.png
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001947106777.png b/docs/modelarts/umn/figure/en-us_image_0000001947106777.png
new file mode 100644
index 00000000..2f2a9624
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001947106777.png differ
diff --git a/docs/modelarts/umn/en-us_image_0000001805380534.png b/docs/modelarts/umn/figure/en-us_image_0000001947141221.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001805380534.png
rename to docs/modelarts/umn/figure/en-us_image_0000001947141221.png
diff --git a/docs/modelarts/umn/en-us_image_0000001852057785.png b/docs/modelarts/umn/figure/en-us_image_0000001947141321.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001852057785.png
rename to docs/modelarts/umn/figure/en-us_image_0000001947141321.png
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001947150553.png b/docs/modelarts/umn/figure/en-us_image_0000001947150553.png
new file mode 100644
index 00000000..340aa338
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001947150553.png differ
diff --git a/docs/modelarts/umn/en-us_image_0000001805391738.png b/docs/modelarts/umn/figure/en-us_image_0000001947156505.png
similarity index 100%
rename from docs/modelarts/umn/en-us_image_0000001805391738.png
rename to docs/modelarts/umn/figure/en-us_image_0000001947156505.png
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001948721605.png b/docs/modelarts/umn/figure/en-us_image_0000001948721605.png
new file mode 100644
index 00000000..ac841993
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001948721605.png differ
diff --git a/docs/modelarts/umn/figure/en-us_image_0000001948721993.png b/docs/modelarts/umn/figure/en-us_image_0000001948721993.png
new file mode 100644
index 00000000..16063e44
Binary files /dev/null and b/docs/modelarts/umn/figure/en-us_image_0000001948721993.png differ
diff --git a/docs/modelarts/umn/inference-modelarts-0001.html b/docs/modelarts/umn/inference-modelarts-0001.html
new file mode 100644
index 00000000..8dba84ee
--- /dev/null
+++ b/docs/modelarts/umn/inference-modelarts-0001.html
@@ -0,0 +1,25 @@
+
+
+
Introduction to Inference
+
After an AI model is developed, you can use it to create an AI application and quickly deploy the application as an inference service. The AI inference capabilities can be integrated into your IT platform by calling APIs.
+
Figure 1 Inference
+
- Develop a model: Models can be developed in ModelArts or your local development environment. A locally developed model must be uploaded to OBS.
- Create an AI application: Import the model file and inference file to the ModelArts model repository and manage them by version. Use these files to build an executable AI application.
- Deploy as a service: Deploy the AI application as a container instance in the resource pool and register inference APIs that can be accessed externally.
- Perform inference: Add the function of calling the inference APIs to your application to integrate AI inference into the service process.
+
Deploying an AI Application as a Service
After an AI application is created, you can deploy it as a service on the
Deploy page. ModelArts supports the following deployment types:
- Real-time service
Deploy an AI application as a web service with real-time test UI and monitoring supported.
+ - Batch service
Deploy an AI application as a batch service that performs inference on batch data and automatically stops after data processing is complete.
+
+
+
+
+
+
+
+
\ No newline at end of file
diff --git a/docs/modelarts/umn/inference-modelarts-0002.html b/docs/modelarts/umn/inference-modelarts-0002.html
new file mode 100644
index 00000000..993bb577
--- /dev/null
+++ b/docs/modelarts/umn/inference-modelarts-0002.html
@@ -0,0 +1,21 @@
+
+
+
Managing AI Applications
+
+
+
diff --git a/docs/modelarts/umn/inference-modelarts-0003.html b/docs/modelarts/umn/inference-modelarts-0003.html
new file mode 100644
index 00000000..603d5739
--- /dev/null
+++ b/docs/modelarts/umn/inference-modelarts-0003.html
@@ -0,0 +1,72 @@
+
+
+
Introduction to AI Application Management
+
AI development and optimization require frequent iterations and debugging. Modifications in datasets, training code, or parameters affect the quality of models. If the metadata of the development process cannot be centrally managed, the optimal model may fail to be reproduced.
+
ModelArts AI application management allows you to import meta models from training jobs, OBS, and container images. In this way, you can centrally manage all iterated and debugged AI applications.
+
Constraints
- In an ExeML project, after a model is deployed, the model is automatically uploaded to the AI application management list. However, AI applications generated by ExeML cannot be downloaded and can be used only for deployment and rollout.
+
+
Scenarios for Creating AI Applications
- Imported from a training job: Create a training job in ModelArts and train a model. After obtaining a satisfactory model, use it to create an AI application and deploy the application as services.
- Imported from OBS: If you use a mainstream framework to develop and train a model locally, you can upload the model to an OBS bucket based on the model package specifications, import the model from OBS to ModelArts, and use the model to create an AI application for service deployment.
- Imported from a container image: If an AI engine is not supported by ModelArts, you can use it to build a model, import the model to ModelArts as a custom image, use the image to create an AI application, and deploy the AI application as services.
+
+
Functions of AI Application Management
+
Table 1 Functions of AI application managementSupported Function
+ |
+Description
+ |
+
|---|
+
+Creating an AI Application
+ |
+Import the trained models to ModelArts and create AI applications for centralized management. The following provides the operation guide for each method of importing models.
+
+ |
+
+Viewing Details About an AI Application
+ |
+After an AI application is created, you can view its information on the details page.
+ |
+
+Managing AI Applications
+ |
+To facilitate traceback and model tuning, ModelArts provides the AI application version management function. You can manage AI applications by version.
+ |
+
+
+
+
+
+
Supported AI Engines for ModelArts Inference
If you import a model from a template or OBS to create an AI application, the following AI engines and versions are supported.
+
- Runtime environments marked with recommended are unified runtime images, which will be used as mainstream base inference images.
- Images of the old version will be discontinued. Use unified images.
- The base images to be removed are no longer maintained.
- Naming a unified runtime image: <AI engine name and version> - <Hardware and version: CPU, CUDA, or CANN> - <Python version> - <OS version> - <CPU architecture>
+
+
+
Table 2 Supported AI engines and their runtimeEngine
+ |
+Runtime
+ |
+
|---|
+
+TensorFlow
+ |
+tensorflow_1.15.0-cann_6.3.0-py_3.7-euler_2.8.3-aarch64
+ |
+
+MindSpore
+ |
+mindspore_2.0.0-cann_6.3.0-py_3.7-euler_2.8.3-aarch64
+ |
+
+PyTorch
+ |
+pytorch_1.11.0-cann_6.3.0-py_3.7-euler_2.8.3-aarch64
+ |
+
+
+
+
+
+
+
+
diff --git a/docs/modelarts/umn/inference-modelarts-0004.html b/docs/modelarts/umn/inference-modelarts-0004.html
new file mode 100644
index 00000000..f3967d57
--- /dev/null
+++ b/docs/modelarts/umn/inference-modelarts-0004.html
@@ -0,0 +1,20 @@
+
+
+
Creating an AI Application
+
+
+
diff --git a/docs/modelarts/umn/inference-modelarts-0005.html b/docs/modelarts/umn/inference-modelarts-0005.html
new file mode 100644
index 00000000..beb8bb7d
--- /dev/null
+++ b/docs/modelarts/umn/inference-modelarts-0005.html
@@ -0,0 +1,135 @@
+
+
+
Viewing Details About an AI Application
+
After an AI application is created, you can view its information on the details page.
- Log in to the ModelArts management console. In the navigation pane on the left, choose AI Application Management > AI Applications. The AI Applications page is displayed.
- Click the name of the target AI application. The application details page is displayed.
On the application details page, you can view the basic information and model precision of the AI application, and switch tab pages to view more information.
+
+Table 1 Basic information about an AI applicationParameter
+ |
+Description
+ |
+
|---|
+
+Name
+ |
+Name of an AI application
+ |
+
+Status
+ |
+Status of an AI application
+ |
+
+Version
+ |
+Current version of an AI application
+ |
+
+ID
+ |
+ID of an AI application
+ |
+
+Size
+ |
+Size of an AI application
+ |
+
+Runtime Environment
+ |
+Runtime environment on which the meta model depends
+ |
+
+Meta Model Source
+ |
+Path to the meta model
+ |
+
+AI Engine
+ |
+AI engine used by an AI application
+ |
+
+Deployment Type
+ |
+Types of the services that an AI application can be deployed
+ |
+
+Model Source
+ |
+Source of a model, which can be ExeML, built-in algorithm, or custom algorithm
+ |
+
+Inference Code
+ |
+Path to the inference code
+ |
+
+Description
+ |
+Click the edit button to add the description of an AI application.
+ |
+
+AI Application Description
+ |
+Description document added during the creation of an AI application
+ |
+
+Associated Training Job
+ |
+Associated training job if the meta model comes from a training job. Click the training job name to go to its details page.
+ |
+
+
+
+
Table 2 Details page of an AI applicationParameter
+ |
+Description
+ |
+
|---|
+
+Model Precision
+ |
+Model recall, precision, accuracy, and F1 score of an AI application
+ |
+
+Parameter Configuration
+ |
+API configuration, input parameters, and output parameters of an AI application
+ |
+
+Runtime Dependency
+ |
+Model dependency on the environment. If creating a job failed, edit the runtime dependency. After the modification is saved, the system will automatically use the original image to create the job again.
+ |
+
+Events
+ |
+The progress of key operations during AI application creation
+Events are stored for three months and will be automatically cleared then.
+For details about the event types and event information of AI applications, see What "Are the Events and Their Types for an AI Application?" in FAQs.
+ |
+
+Constraint
+ |
+Constraints on deployment, such as the API request mode, deployed system architecture, and supported acceleration card types
+ |
+
+Associated Services
+ |
+The list of services that an AI application was deployed. Click a service name to go to the service details page.
+ |
+
+
+
+
+
+
+
+
diff --git a/docs/modelarts/umn/inference-modelarts-0006.html b/docs/modelarts/umn/inference-modelarts-0006.html
new file mode 100644
index 00000000..ec0e6566
--- /dev/null
+++ b/docs/modelarts/umn/inference-modelarts-0006.html
@@ -0,0 +1,89 @@
+
+
+
Importing a Meta Model from a Training Job
+
You can create a training job in ModelArts to obtain a satisfactory model. Then, you can import the model to AI Application Management for centralized management. In addition, you can quickly deploy the model as a service.
+
Constraints
- A model generated from a training job that uses subscribed algorithms can be directly imported to ModelArts without the need to use the inference code or configuration file.
- ModelArts of the Arm architecture does not support model import from training.
+
+
Prerequisites
- The training job has been successfully executed, and the model has been stored in the OBS directory where the training output is stored (the input parameter is train_url).
- If a model is generated from a training job that uses a frequently-used framework or custom image, upload the inference code and configuration file to the storage directory of the model by referring to Introduction.
- The OBS directory you use and ModelArts are in the same region.
+
+
Creating an AI Application
- Log in to the ModelArts management console and choose AI Application Management > AI Applications in the left navigation pane. The AI Applications page is displayed.
- Click Create in the upper left corner.
- On the displayed page, configure parameters.
- Set basic information about the AI application. For details about the parameters, see Table 1.
+
Table 1 Parameters of basic AI application informationParameter
+ |
+Description
+ |
+
|---|
+
+Name
+ |
+Application name. The value can contain 1 to 64 visible characters. Only letters, digits, hyphens (-), and underscores (_) are allowed.
+ |
+
+Version
+ |
+Version of the AI application to be created. For the first creation, the default value is 0.0.1.
+ NOTE: After an AI application is created, you can create new versions using different meta models for optimization.
+
+ |
+
+Description
+ |
+Brief description of an AI application
+ |
+
+
+
+
- Select the meta model source and set related parameters. Set Meta Model Source to Training job. For details about the parameters, see Table 2.
+
Table 2 Parameters of the meta model source Parameter
+ |
+Description
+ |
+
|---|
+
+Meta Model Source
+ |
+Select Training job.
+Choose a completed training job under the current account from the Training Job drop-down list.
+ |
+
+AI Engine
+ |
+Inference engine used by the meta model. The engine is automatically matched based on the training job you select.
+ |
+
+Inference Code
+ |
+Set inference code for an AI application. The code is used to customize the inference processing logic. Display the inference code URL. You can copy this URL directly.
+ |
+
+Runtime Dependency
+ |
+List the dependencies of the selected model in the environment.
+ |
+
+AI Application Description
+ |
+Provide AI application descriptions to help other AI application developers better understand and use your applications. Click Add AI Application Description and set the document name and URL. A maximum of three AI application descriptions are supported.
+ |
+
+Deployment Type
+ |
+Select the service types that the application can be deployed. When deploying a service, only the service types selected here are available. For example, if you only select Real-Time Services here, you can only deploy the AI application as a real-time service after it is created.
+ |
+
+
+
+
- Confirm the configurations and click Create now. The AI application is created.
In the AI application list, you can view the created AI application and its version. When the status changes to Normal, the AI application is successfully created. On this page, you can perform such operations as creating new versions, quickly deploying AI applications, and publishing AI applications.
+
+
+
+
Follow-Up Procedure
Deploying an AI Application as a Service: In the AI application list, click the down arrow on the left of an AI application name to check all versions of the AI application. Locate the row that contains the target version, click Deploy in the Operation column, and select a deployment type from the drop-down list box. The AI application can be deployed in a deployment type selected during AI application creation.
+
+
+
+
diff --git a/docs/modelarts/umn/inference-modelarts-0008.html b/docs/modelarts/umn/inference-modelarts-0008.html
new file mode 100644
index 00000000..005efdde
--- /dev/null
+++ b/docs/modelarts/umn/inference-modelarts-0008.html
@@ -0,0 +1,100 @@
+
+
+
Importing a Meta Model from OBS
+
If a model is developed and trained using a mainstream AI engine, import the model to ModelArts and use the model to create an AI application. In this way, the AI applications can be centrally managed on ModelArts.
+
+
Prerequisites
- The model has been developed and trained, and the type and version of the AI engine used by the model are supported by ModelArts. For details, see Supported AI Engines for ModelArts Inference.
- The trained model package, inference code, and configuration file have been uploaded to OBS.
- The OBS directory you use and ModelArts are in the same region.
+
+
Creating an AI Application
- Log in to the ModelArts console, and choose AI Application Management > AI Applications from the navigation pane. The AI Applications page is displayed.
- Click Create in the upper left corner.
- On the displayed page, configure parameters.
- Enter basic information about the AI application. For details, see Table 1.
+
Table 1 Basic informationParameter
+ |
+Description
+ |
+
|---|
+
+Name
+ |
+Name of the AI application. The value can contain 1 to 64 visible characters. Only letters, digits, hyphens (-), and underscores (_) are allowed.
+ |
+
+Version
+ |
+Version of the AI application. The default value is 0.0.1 for the first import.
+ NOTE: After an AI application is created, you can create new versions using different meta models for optimization.
+
+ |
+
+Description
+ |
+Brief description of the AI application.
+ |
+
+
+
+
- Select the meta model source and configure related parameters. Set Meta Model Source to OBS. For details about the parameters, see Table 2.
To import a meta model from OBS, edit the inference code and configuration file by following model package specifications and place the inference code and configuration file in the model folder storing the meta model. If the selected directory does not comply with the model package specifications, the AI application cannot be created.
+
+Table 2 Meta model source parametersParameter
+ |
+Description
+ |
+
|---|
+
+Meta Model
+ |
+OBS path for storing the meta model.
+The OBS path cannot contain spaces. Otherwise, the creation of the AI application will fail.
+ |
+
+AI Engine
+ |
+AI engine, which is automatically set according to the model storage path you select, used by the meta model.
+If the AI Engine is Custom, configure the following parameters:
+- Container API: Protocol and port number for starting the model. The default request protocol and port number are HTTPS and 8080, respectively.
- Health Check: Health check on the model. This parameter is configurable only when the health check API is configured in the custom image. Otherwise, the AI application deployment will fail.
- Check Mode: Select HTTP request or Command.
- Health Check URL: Enter the health check URL, which defaults to /health. This parameter is displayed when Check Mode is set to HTTP request.
- Health Check Command: Enter the health check command. This parameter is displayed when Check Mode is set to Command.
- Health Check Period: Enter an integer ranging from 1 to 2147483647. The unit is second. The default value is 5.
- Delay: Set a delay for the health check to occur after the instance has started. Enter an integer ranging from 0 to 2147483647. The unit is second. The default value is 12.
- Maximum Failures: Enter an integer ranging from 1 to 2147483647. If the service fails the specified number of consecutive health checks during startup, it will enter the abnormal state. If the service fails the specified number of consecutive health checks during operation, it will enter the alarm state. The default value is 12.
+
+ |
+
+Runtime Dependency
+ |
+Dependencies that the selected model has on the environment.
+ |
+
+AI Application Description
+ |
+AI application descriptions to help other developers better understand and use your application. Click Add AI Application Description and enter the document name and URL. You can add up to three descriptions.
+ |
+
+Configuration File
+ |
+The system associates the configuration file stored in OBS by default. After enabling this function, you can review and edit the model configuration file.
+ NOTE: This function is to be discontinued. After that, you can modify the model configuration by setting AI Engine, Runtime Dependency, and API Configuration.
+
+ |
+
+Deployment Type
+ |
+Choose the service types for application deployment. The service types you select will be the only options available for deployment. For instance, selecting Real-Time Services means the AI application can only be deployed as real-time services.
+ |
+
+API Configuration
+ |
+You can enable it to edit RESTful APIs to define the AI application input and output formats. The API configuration must comply with ModelArts specifications. For details, see apis parameters in Specifications for Editing a Model Configuration File. Code Example of apis Parameters shows an example.
+ |
+
+
+
+
- Check the information and click Create now.
In the AI application list, you can view the created AI application and its version. When the status changes to Normal, the AI application is created. On this page, you can perform such operations as creating versions, publishing AI applications, and deploying services.
+
+
+
+
Follow-Up Operations
Deploying an AI Application as a Service: In the AI application list, click the down arrow on the left of an AI application name to check all versions of the AI application. Locate the row that contains the target version, click Deploy in the Operation column, and select a deployment type from the drop-down list. The AI application can be deployed as a deployment type selected during AI application creation.
+
+
+
+
diff --git a/docs/modelarts/umn/inference-modelarts-0009.html b/docs/modelarts/umn/inference-modelarts-0009.html
new file mode 100644
index 00000000..ed0f329c
--- /dev/null
+++ b/docs/modelarts/umn/inference-modelarts-0009.html
@@ -0,0 +1,100 @@
+
+
+
Importing a Meta Model from a Container Image
+
For AI engines that are not supported by ModelArts, you can import the models you compile to ModelArts from custom images.
+
+
Prerequisites
The OBS directory you use and ModelArts are in the same region.
+
+
Creating an AI Application
- Log in to the ModelArts management console, and choose AI Application Management > AI Applications in the left navigation pane. The AI Applications page is displayed.
- Click Create in the upper left corner.
- On the displayed page, set the parameters.
- Set basic information about the AI application. For details about the parameters, see Table 1.
+
Table 1 Parameters of basic AI application informationParameter
+ |
+Description
+ |
+
|---|
+
+Name
+ |
+Application name. The value can contain 1 to 64 visible characters. Only letters, digits, hyphens (-), and underscores (_) are allowed.
+ |
+
+Version
+ |
+Version of the AI application to be created. For the first import, the default value is 0.0.1.
+ NOTE: After an AI application is created, you can create new versions using different meta models for optimization.
+
+ |
+
+Description
+ |
+Brief description of an AI application
+ |
+
+
+
+
- Select the meta model source and set related parameters. Set Meta Model Source to Container image. For details about the parameters, see Table 2.
+
Table 2 Parameters of the meta model source Parameter
+ |
+Description
+ |
+
|---|
+
+Container Image Path
+ |
+Click  to import the model image from the container image. The model is of the Image type, and you do not need to use swr_location in the configuration file to specify the image location.
+For details about how to create a custom image, see Custom Image Specifications for Creating AI Applications.
+ NOTE: The model image you select will be shared with the system administrator, so ensure you have the permission to share the image (images shared with other accounts are unsupported). When you deploy a service, ModelArts deploys the image as an inference service. Ensure that your image can be properly started and provide an inference interface.
+
+ |
+
+Image Replication
+ |
+Indicates whether to copy the model image in the container image to ModelArts.
+- When this function is disabled, the model image is not copied, AI applications can be created quickly, but modifying or deleting images in the source directory of SWR may affect service deployment.
- When this function is enabled, the model image is copied, AI applications cannot be created quickly, but you can modify or delete images in the source directory of SWR as that would not affect service deployment.
+ |
+
+Health Check
+ |
+Health check on an AI application. This parameter is configurable only when the health check API is configured in the custom image. Otherwise, the AI application deployment will fail.
+- Check Mode: Select HTTP request or Command.
- Health Check URL: This parameter is displayed when Check Mode is set to HTTP request. Enter the health check URL. The default value is /health.
- Health Check Command: This parameter is displayed when Check Mode is set to Command. Enter the health check command.
- Health Check Period: Enter an integer ranging from 1 to 2147483647. The unit is second. The default value is 5.
- Delay: specifies the delay for performing the health check after the instance is started. Enter an integer ranging from 0 to 2147483647. The unit is second and cannot be empty. The default value is 12.
- Maximum Failures: Enter an integer ranging from 1 to 2147483647. During service startup, if the number of consecutive health check failures reaches the specified value, the service will be abnormal. During service running, if the number of consecutive health check failures reaches the specified value, the service will enter the alarm status. The default value is 12.
+ NOTE: If health check is configured for an AI application, the deployed services using this AI application will stop 3 minutes after receiving the stop instruction.
+
+ |
+
+AI Application Description
+ |
+Provide AI application descriptions to help other AI application developers better understand and use your applications. Click Add AI Application Description and set the Document name and URL. You can add up to three AI application descriptions.
+ |
+
+Deployment Type
+ |
+Select the service types that the application can be deployed. When deploying a service, only the service types selected here are available. For example, if you only select Real-time services here, you can only deploy the AI application as a real-time service after it is created.
+ |
+
+Start command
+ |
+customizable start command of a model
+ |
+
+Apis
+ |
+When you enable this function, you can edit RESTful APIs to define the AI application input and output formats. The API configuration must comply with ModelArts specifications. For details, see apis parameters in Specifications for Editing a Model Configuration File. Code Example of apis Parameters shows an example.
+ |
+
+
+
+
- Check the information and click Next. The AI application is created.
In the AI application list, you can view the created AI application and its version. When the status changes to Normal, the AI application is successfully created. On this page, you can perform such operations as creating new versions, quickly deploying AI applications, and publishing AI applications.
+
+
+
+
Follow-Up Procedure
Deploying an AI Application as a Service: In the AI application list, click the down arrow on the left of an AI application name to check all versions of the AI application. Locate the row that contains the target version, click Deploy in the Operation column, and select a deployment type from the drop-down list box. The AI application can be deployed in a deployment type selected during AI application creation.
+
+
+
+
diff --git a/docs/modelarts/umn/inference-modelarts-0013.html b/docs/modelarts/umn/inference-modelarts-0013.html
new file mode 100644
index 00000000..bc3cfedf
--- /dev/null
+++ b/docs/modelarts/umn/inference-modelarts-0013.html
@@ -0,0 +1,23 @@
+
+
+
Managing AI Applications
+
To facilitate source tracing and repeated AI application tuning, ModelArts provides the AI application version management function. You can manage models based on versions.
+
Prerequisites
An AI application has been created in ModelArts.
+
+
Creating a New Version
On the AI Application Management > AI Applications page, click Create Version in the Operation column of the target AI application. On the Create Version page, set the parameters. For details, see Creating an AI Application. Click Create now.
+
+
Deleting a Version
On the AI Application Management > AI Applications page, click the downward arrow on the left of the AI application name to expand an application version list. In the application version list, click Delete in the Operation column to delete the corresponding version.
+
If a service has been deployed for the AI application version, you need to delete the associated service before deleting the AI application version. A deleted version cannot be recovered. Exercise caution when performing this operation.
+
+
+
Deleting an AI Application
In the navigation pane, choose AI Application Management > AI Applications. On the AI Applications page, click Delete in the Operation column to delete the target AI application.
+
If a service has been deployed for the AI application version, you need to delete the associated service before deleting the AI application version. A deleted AI application cannot be recovered. Exercise caution when performing this operation.
+
+
+
+
+
diff --git a/docs/modelarts/umn/inference-modelarts-0016.html b/docs/modelarts/umn/inference-modelarts-0016.html
new file mode 100644
index 00000000..f686fb62
--- /dev/null
+++ b/docs/modelarts/umn/inference-modelarts-0016.html
@@ -0,0 +1,21 @@
+
+
+
Deploying AI Applications as Real-Time Services
+
+
+
diff --git a/docs/modelarts/umn/inference-modelarts-0018.html b/docs/modelarts/umn/inference-modelarts-0018.html
new file mode 100644
index 00000000..734a73cb
--- /dev/null
+++ b/docs/modelarts/umn/inference-modelarts-0018.html
@@ -0,0 +1,148 @@
+
+
+
Deploying as a Real-Time Service
+
After an AI application is prepared, you can deploy the AI application as a real-time service and predict and call the service.
+
Constraints
A maximum of 20 real-time services can be deployed by a user.
+
+
Prerequisites
- Data has been prepared. Specifically, you have created an AI application in the Normal state in ModelArts.
+
+
Note
Real-time services deployed using the public resource pool also occupy quota resources when the services are Abnormal or Stopped. If the quota is insufficient and no more services can be deployed, delete some abnormal services to release resources.
+
Quota calculation:
+
- If a dedicated resource pool is used to deploy real-time services, the quota is not decreased. The quota is increased or decreased only when the dedicated pool is created, modified, or deleted.
- When a shared resource pool is used to deploy a real-time service, the quota will be increased or decreased when you create, change the number of, or delete instances.
+
Metering calculation:
+
- If a real-time service is deployed using a dedicated pool, only the data of the dedicated pool to which the service belongs is metered.
- When a shared pool is used to deploy a real-time service, the specifications used by the service will be metered.
+
+
Procedure
- Log in to the ModelArts management console. In the left navigation pane, choose Service Deployment > Real-Time Services. By default, the system switches to the Real-Time Services page.
- In the real-time service list, click Deploy in the upper left corner. The Deploy page is displayed.
- Set parameters for a real-time service.
- Set basic information about model deployment. For details about the parameters, see Table 1.
+
Table 1 Basic parameters of model deploymentParameter
+ |
+Description
+ |
+
|---|
+
+Name
+ |
+Name of the real-time service. Set this parameter as prompted.
+ |
+
+Description
+ |
+Brief description of the real-time service.
+ |
+
+
+
+
- Enter key information including the resource pool and AI application configurations. For details, see Table 2.
+
Table 2 ParametersParameter
+ |
+Sub-Parameter
+ |
+Description
+ |
+
|---|
+
+Resource Pool
+ |
+Public resource pools
+ |
+Instances in the public resource pool can be of the CPU or GPU type.
+ |
+
+Resource Pool
+ |
+Dedicated resource pools
+ |
+Select a specification from the dedicated resource pool specifications.
+ NOTE: - The data of old-version dedicated resource pools will be gradually migrated to the new-version dedicated resource pools.
- For new users and the existing users who have migrated data from old-version dedicated resource pools to new ones, there is only one entry to new-version dedicated resource pools on the ModelArts management console.
- For the existing users who have not migrated data from old-version dedicated resource pools to new ones, there are two entries to dedicated resource pools on the ModelArts management console, where the entry marked with New is to the new version.
+ For more details about the new-version dedicated resource pools, see Comprehensive Upgrades to ModelArts Resource Pool Management Functions.
+
+ |
+
+Multi-Pool Load Balancing
+ |
+N/A
+ |
+After this function is enabled, the service will be deployed in two dedicated resource pools, and service traffic will be evenly distributed among the pools through a load balancer. This minimizes the impact on the service after one resource pool fails, improving service reliability.
+After this function is enabled, the number of compute nodes must be a multiple of 2 and the minimum value is 2.
+ NOTE: - Multi-pool load balancing is supported only when a dedicated resource pool is selected.
- Multi-pool load balancing requires that the compute nodes in the two resource pools have the same specifications.
- Both resource pools must be old or new resource pools.
+
+ |
+
+AI Application and Configuration
+ |
+AI Application Source
+ |
+Select My AI Applications based on your requirements.
+ |
+
+AI Application and Version
+ |
+Select the AI application and version that are in the Normal state.
+ |
+
+Traffic Ratio (%)
+ |
+Set the traffic proportion of the current instance node. Service calling requests are allocated to the current version based on this proportion.
+If you deploy only one version of an AI application, set this parameter to 100%. If you select multiple versions for gated launch, ensure that the sum of the traffic ratios of multiple versions is 100%.
+ |
+
+Specifications
+ |
+Select available specifications based on the list displayed on the console. The specifications in gray cannot be used in the current environment.
+If specifications in the public resource pools are unavailable, no public resource pool is available in the current environment. In this case, use a dedicated resource pool or contact the administrator to create a public resource pool.
+ NOTE: When the selected flavor is used to deploy the service, necessary system consumption is generated. Therefore, the resources actually occupied by the service are slightly greater than the selected flavor.
+
+ |
+
+Compute Nodes
+ |
+Set the number of instances for the current AI application version. If you set Instances to 1, the standalone computing mode is used. If you set Instances to a value greater than 1, the distributed computing mode is used. Select a computing mode based on the actual requirements.
+ |
+
+Environment Variable
+ |
+Set environment variables and inject them to the pod. To ensure data security, do not enter sensitive information in environment variables.
+ |
+
+Timeout
+ |
+Timeout of a single model, including both the deployment and startup time. The default value is 20 minutes. The value must range from 3 to 120.
+ |
+
+WebSocket
+ |
+Whether to deploy a real-time service as a WebSocket service.
+ NOTE: - This function is supported only if the AI application is WebSocket-compliant and comes from a container image.
- After this function is enabled, Traffic Limit and Data Collection cannot be set.
- This parameter cannot be changed after the service is deployed.
+
+ |
+
+Add AI Application Version and Configuration
+ |
+If the selected AI application has multiple versions, you can add multiple versions and configure a traffic ratio. You can use gray launch to smoothly upgrade the AI application version.
+ NOTE: Free compute specifications do not support the gray launch of multiple versions.
+
+ |
+
+Data Collection
+ |
+N/A
+ |
+This function is disabled by default. When enabled, it collects and stores data generated when a real-time service is called based on configured rules.
+ |
+
+
+
+
+ - After confirming the entered information, complete service deployment as prompted. Generally, service deployment jobs run for a period of time, which may be several minutes or tens of minutes depending on the amount of your selected data and resources.
After a real-time service is deployed, it is started immediately.
+
+In the real-time service list, after the status of the newly deployed service changes from Deploying to Running, the service is deployed successfully.
+
+
+
+
+
diff --git a/docs/modelarts/umn/inference-modelarts-0019.html b/docs/modelarts/umn/inference-modelarts-0019.html
new file mode 100644
index 00000000..3c786f94
--- /dev/null
+++ b/docs/modelarts/umn/inference-modelarts-0019.html
@@ -0,0 +1,109 @@
+
+
+
Viewing Service Details
+
After an AI application is deployed as a real-time service, you can access the service page to view its details.
+
- Log in to the ModelArts management console and choose Service Deployment > Real-Time Services.
- On the Real-Time Services page, click the name of the target service. The service details page is displayed.
You can view the service name, status, and other information. For details, see Table 1.
+
+Table 1 real-time service parametersParameter
+ |
+Description
+ |
+
|---|
+
+Name
+ |
+Name of the real-time service.
+ |
+
+Status
+ |
+Status of the real-time service.
+ |
+
+Source
+ |
+AI application source of the real-time service.
+ |
+
+Service ID
+ |
+Real-time service ID
+ |
+
+Failed Calls/Total Calls
+ |
+Number of service calls, which is counted from the time when the service was created.
+If the number of AI applications is changed or a service is invoked when an AI application is not ready, the number of calls is not counted.
+ |
+
+Description
+ |
+Service description, which can be edited after you click the edit button on the right side.
+ |
+
+Data Collection
+ |
+Enable this option to store the data generated when the real-time service is invoked to a specified OBS path.
+ |
+
+WebSocket
+ |
+Whether to upgrade to the WebSocket service.
+ |
+
+
+
+
- You can switch between tabs on the details page of a real-time service to view more details. For details, see Table 2.
+
Table 2 Details of a real-time serviceParameter
+ |
+Description
+ |
+
|---|
+
+Usage Guides
+ |
+Displays the API URL, AI application information, input parameters, and output parameters. You can click  to copy the API URL to call the service.
+ |
+
+Prediction
+ |
+Performs a prediction test on the real-time service. For details, see Testing the Deployed Service.
+ |
+
+Configuration Updates
+ |
+Displays Existing Configuration and Historical Updates.
+- Existing Configuration: includes the AI application name, version, status, compute node flavor, the number of compute nodes, and traffic ratio.
- Historical Updates: displays historical AI application information.
+ |
+
+Monitoring
+ |
+Displays Resource Usage and AI Application Calls.
+- Resource Usage: includes the used and available CPU, memory, and GPU resources.
- AI Application Calls: indicates the number of AI application calls. The statistics collection starts after the AI application status changes to Ready. (This parameter is not displayed for WebSocket services.)
+ |
+
+Event
+ |
+Displays key operations during service use, such as the service deployment progress, detailed causes of deployment exceptions, and time points when a service is started, stopped, or modified.
+Events are saved for one month and will be automatically cleared then.
+For details about the event types and event information of services, see What Are the Events and Their Types for a Service?
+ |
+
+Logs
+ |
+Displays the log information about each AI application in the service. You can view logs generated in the latest 5 minutes, latest 30 minutes, latest 1 hour, and user-defined time segment.
+You can select the start time and end time when defining the time segment.
+ |
+
+
+
+
+
+
+
diff --git a/docs/modelarts/umn/inference-modelarts-0020.html b/docs/modelarts/umn/inference-modelarts-0020.html
new file mode 100644
index 00000000..8cd6ddc9
--- /dev/null
+++ b/docs/modelarts/umn/inference-modelarts-0020.html
@@ -0,0 +1,24 @@
+
+
+
Testing the Deployed Service
+
After an AI application is deployed as a real-time service, you can debug code or add files for testing on the Prediction tab page. Based on the input request (JSON text or file) defined by the AI application, the service can be tested in either of the following ways:
+
- JSON Text Prediction: If the input type of the AI application of the deployed service is JSON text, that is, the input does not contain files, you can enter the JSON code on the Prediction tab page for service testing.
- File Prediction: If the input type of the AI application of the deployed service is file, including images, audios, and videos, you can add images on the Prediction tab page for service testing.
+
- If the input type is image, the size of a single image must be less than 8 MB.
- The following image types are supported: png, psd, jpg, jpeg, bmp, gif, webp, psd, svg, and tiff.
- This function is used for commissioning. In actual production, you are advised to call APIs. Access the real-time service through token-based authentication.
+
+
Input Parameters
For the service that you have deployed, you can learn about its input parameters of the service, that is, the input request type mentioned above, on the Usage Guides tab page of the service details page.
+
The input parameters displayed on the Usage Guides tab page depend on the AI application source that you select.
+
- If your metamodel comes from ExeML or a built-in algorithm, the input and output parameters are defined by ModelArts. For details, see the Usage Guides tab page. On the Prediction tab page, enter the corresponding JSON text or file for service testing.
- If you use a custom meta model with the inference code and configuration file compiled by yourself (Specifications for Writing the Model Configuration File), the Usage Guides tab page only visualizes your data. The following figure shows the mapping between the input parameters displayed on the Usage Guides tab page and the configuration file.
+
+
JSON Text Prediction
- Log in to the ModelArts management console and choose Service Deployment > Real-Time Services.
- On the Real-Time Services page, click the name of the target service. The service details page is displayed. On the Prediction tab page, enter the prediction code and click Predict to perform prediction.
+
+
+
File Prediction
- Log in to the ModelArts management console and choose Service Deployment > Real-Time Services.
- On the Real-Time Services page, click the name of the target service. The service details page is displayed. On the Prediction tab page, click Upload and select a test file. After the file is uploaded successfully, click Predict to perform a prediction test.
+
+
+
+
+
diff --git a/docs/modelarts/umn/inference-modelarts-0039.html b/docs/modelarts/umn/inference-modelarts-0039.html
new file mode 100644
index 00000000..8bcc195b
--- /dev/null
+++ b/docs/modelarts/umn/inference-modelarts-0039.html
@@ -0,0 +1,17 @@
+
+
+
Deploying AI Applications as Batch Services
+
+
+
diff --git a/docs/modelarts/umn/inference-modelarts-0040.html b/docs/modelarts/umn/inference-modelarts-0040.html
new file mode 100644
index 00000000..2bee5006
--- /dev/null
+++ b/docs/modelarts/umn/inference-modelarts-0040.html
@@ -0,0 +1,247 @@
+
+
+
Deploying as a Batch Service
+
After an AI application is prepared, you can deploy it as a batch service. The Service Deployment > Batch Services page lists all batch services. You can enter a service name in the search box in the upper right corner and click 
to search for the service.
+
Prerequisites
- A ModelArts application in the Normal state is available.
- Data to be batch processed is ready and has been upload to an OBS directory.
- At least one empty folder has been created in OBS for storing the output.
+
+
Context
- A maximum of 1,000 batch services can be created.
- Based on the input request (JSON or file) defined by the AI application, different parameters are entered. If the AI application input is a JSON file, a configuration file is required to generate a mapping file. If the AI application input is a file, no mapping file is required.
- Batch services can only be deployed in a public resource pool, but not a dedicated resource pool.
+
+
Procedure
- Log in to the ModelArts management console. In the left navigation pane, choose Service Deployment > Batch Services. By default, the Batch Services page is displayed.
- In the batch service list, click Deploy in the upper left corner. The Deploy page is displayed.
- Set parameters for a batch service.
- Set the basic information, including Name and Description. The name is generated by default, for example, service-bc0d. You can specify Name and Description according to actual requirements.
- Set other parameters, including AI application configurations. For details, see Table 1.
+
Table 1 ParametersParameter
+ |
+Description
+ |
+
|---|
+
+AI Application Source
+ |
+Select My AI Applications based on your requirements.
+ |
+
+AI Application and Version
+ |
+Select an AI application and version that are running properly.
+ |
+
+Input Path
+ |
+Select the OBS directory where the uploaded data is stored. Select a folder or a .manifest file. For details about the specifications of the .manifest file, see Manifest File Specifications.
+ NOTE: - If the input data is an image, ensure that the size of a single image is less than 10 MB.
- If the input data is in CSV format, ensure that the file size does not exceed 12 MB.
+
+ |
+
+Request Path
+ |
+URI used for calling the AI application interface in a batch service, and also the request path of the AI application service. Its value is obtained from the url field of apis in the AI application configuration file.
+ |
+
+Mapping Relationship
+ |
+If the AI application input is in JSON format, the system automatically generates the mapping based on the configuration file corresponding to the AI application. If the AI application input is other file, mapping is not required.
+Automatically generated mapping file. Enter the field index corresponding to each parameter in the CSV file. The index starts from 0.
+Mapping rule: The mapping rule comes from the input parameter (request) in the model configuration file config.json. When type is set to string, number, integer, or boolean, you are required to set the index parameter. For details about the mapping rule, see Example Mapping.
+The index must be a positive integer starting from 0. If the value of index does not comply with the rule, this parameter is ignored in the request. After the mapping rule is configured, the corresponding CSV data must be separated by commas (,).
+ |
+
+Output Path
+ |
+Select the path for saving the batch prediction result. You can select the empty folder that you create.
+ |
+
+Specifications
+ |
+Select available specifications based on the list displayed on the console. The specifications in gray cannot be used at the current region.
+ |
+
+Compute Nodes
+ |
+Set the number of instances for the current AI application version. If you set the number of nodes to 1, the standalone computing mode is used. If you set the number of nodes to a value greater than 1, the distributed computing mode is used. Select a computing mode based on the actual requirements.
+ |
+
+Environment Variable
+ |
+Set environment variables and inject them to the pod. To ensure data security, do not enter sensitive information in environment variables.
+ |
+
+Timeout
+ |
+Timeout of a single model, including both the deployment and startup time. The default value is 20 minutes. The value must range from 3 to 120.
+ |
+
+
+
+
+ - After setting the parameters, deploy the model as a batch service as prompted. Deploying a service generally requires a period of time, which may be several minutes or tens of minutes depending on the amount of your data and resources.
You can go to the batch service list to view the basic information about the batch service. In the batch service list, after the status of the newly deployed service changes from Deploying to Running, the service is deployed successfully.
+
+
+
Manifest File Specifications
ModelArts batch services support manifest files, which describe data input and output.
+
+
Example input manifest file
+
+
Example output manifest file
+
A manifest file will be generated in the output directory of the batch services.
- Assume that the output path is //test-bucket/test/. The result is stored in the following path:
OBS bucket/directory name
+├── test-bucket
+│ ├── test
+│ │ ├── infer-result-{{task_id}}.manifest
+│ │ ├── infer-result
+│ │ │ ├── 1.jpg_result.txt
+│ │ │ ├── 2.jpg_result.txt
+ - Content of the infer-result-0.manifest file:
{"source": "obs://obs-data-bucket/test/data/1.jpg","result":"SUCCESSFUL","inference-loc": "obs://test-bucket/test/infer-result/1.jpg_result.txt"}
+{"source": "s3://obs-data-bucket/test/data/2.jpg","result":"FAILED","error_message": "Download file failed."}
+{"source ": "https://infers-data.obs.xxx.com:443/xgboosterdata/2.jpg?AccessKeyId=2Q0V0TQ461N26DDL18RB&Expires=1550611914&Signature=wZBttZj5QZrReDhz1uDzwve8GpY%3D&x-obs-security-token=gQpzb3V0aGNoaW5hixvY8V9a1SnsxmGoHYmB1SArYMyqnQT-ZaMSxHvl68kKLAy5feYvLDMNZWxzhBZ6Q-3HcoZMh9gISwQOVBwm4ZytB_m8sg1fL6isU7T3CnoL9jmvDGgT9VBC7dC1EyfSJrUcqfB_N0ykCsfrA1Tt_IQYZFDu_HyqVk-GunUcTVdDfWlCV3TrYcpmznZjliAnYUO89kAwCYGeRZsCsC0ePu4PHMsBvYV9gWmN9AUZIDn1sfRL4voBpwQnp6tnAgHW49y5a6hP2hCAoQ-95SpUriJ434QlymoeKfTHVMKOeZxZea-JxOvevOCGI5CcGehEJaz48sgH81UiHzl21zocNB_hpPfus2jY6KPglEJxMv6Kwmro-ZBXWuSJUDOnSYXI-3ciYjg9-h10b8W3sW1mOTFCWNGoWsd74it7l_5-7UUhoIeyPByO_REwkur2FOJsuMpGlRaPyglZxXm_jfdLFXobYtzZhbul4yWXga6oxTOkfcwykTOYH0NPoPRt5MYGYweOXXxFs3d5w2rd0y7p0QYhyTzIkk5CIz7FlWNapFISL7zdhsl8RfchTqESq94KgkeqatSF_iIvnYMW2r8P8x2k_eb6NJ7U_q5ztMbO9oWEcfr0D2f7n7Bl_nb2HIB_H9tjzKvqwngaimYhBbMRPfibvttW86GiwVP8vrC27FOn39Be9z2hSfJ_8pHej0yMlyNqZ481FQ5vWT_vFV3JHM-7I1ZB0_hIdaHfItm-J69cTfHSEOzt7DGaMIES1o7U3w%3D%3D","result":"SUCCESSFUL","inference-loc": "obs://test-bucket/test/infer-result/2.jpg_result.txt"}
+
+
+
- File format:
- The file name is infer-result-{{task_id}}.manifest, where task_id is the batch task ID, which is unique for a batch service.
- If a large number of files need to be processed, multiple manifest files may be generated with the same suffix .manifest and are distinguished by suffix, for example, infer-result-{{task_id}}_1.manifest.
- The infer-result-{{task_id}} directory is created in the manifest directory to store the file processing result.
- The file content is in JSON format. Each row describes the output result of a piece of input data.
- The file contains multiple fields:
- source: input data description, which is the same as that of the input manifest file
- result: file processing result, which can be SUCCESSFUL or FAILED
- inference-loc: output result path. This field is available when result is SUCCESSFUL. The format is obs://{{Bucket name}}/{Object name}.
- error_message: error information. This field is available when the result is FAILED.
+
+
+
Example Mapping
The following example shows the relationship between the configuration file, mapping rule, CSV data, and inference request.
+
The following uses a file for prediction as an example:
+
1
+ 2
+ 3
+ 4
+ 5
+ 6
+ 7
+ 8
+ 9
+10
+11
+12
+13
+14
+15
+16
+17
+18
+19
+20
+21
+22
+23
+24
+25
+26
+27
+28
+29
+30
+31
+32
+33
+34
+35
+36
+37
+38
+39
+40
+41 | [
+ {
+ "method": "post",
+ "url": "/",
+ "request": {
+ "Content-type": "multipart/form-data",
+ "data": {
+ "type": "object",
+ "properties": {
+ "data": {
+ "type": "object",
+ "properties": {
+ "req_data": {
+ "type": "array",
+ "items": [
+ {
+ "type": "object",
+ "properties": {
+ "input_1": {
+ "type": "number"
+ },
+ "input_2": {
+ "type": "number"
+ },
+ "input_3": {
+ "type": "number"
+ },
+ "input_4": {
+ "type": "number"
+ }
+ }
+ }
+ ]
+ }
+ }
+ }
+ }
+ }
+ }
+ }
+]
+ |
+
+
+
The ModelArts management console automatically resolves the mapping relationship from the configuration file as shown below. When calling a ModelArts API, configure the mapping by following the rule.
+
{
+ "type": "object",
+ "properties": {
+ "data": {
+ "type": "object",
+ "properties": {
+ "req_data": {
+ "type": "array",
+ "items": [
+ {
+ "type": "object",
+ "properties": {
+ "input_1": {
+ "type": "number",
+ "index": 0
+ },
+ "input_2": {
+ "type": "number",
+ "index": 1
+ },
+ "input_3": {
+ "type": "number",
+ "index": 2
+ },
+ "input_4": {
+ "type": "number",
+ "index": 3
+ }
+ }
+ }
+ ]
+ }
+ }
+ }
+ }
+}
+
Multiple pieces of CSV data for inference are separated by commas (,) The following shows an example:
+
5.1,3.5,1.4,0.2
+4.9,3.0,1.4,0.2
+4.7,3.2,1.3,0.2
+
Depending on the defined mapping relationship, the inference request is shown below, whose format is similar to that for real-time services.
+
{
+ "data": {
+ "req_data": [{
+ "input_1": 5.1,
+ "input_2": 3.5,
+ "input_3": 1.4,
+ "input_4": 0.2
+ }]
+ }
+}
+
+
+
+
diff --git a/docs/modelarts/umn/inference-modelarts-0041.html b/docs/modelarts/umn/inference-modelarts-0041.html
new file mode 100644
index 00000000..ee3d5715
--- /dev/null
+++ b/docs/modelarts/umn/inference-modelarts-0041.html
@@ -0,0 +1,17 @@
+
+
+
Viewing the Batch Service Prediction Result
+
When deploying a batch service, you can select the location of the output data directory. You can view the running result of the batch service that is in the Completed status.
+
Procedure
- Log in to the ModelArts management console and choose Service Deployment > Batch Services.
- Click the name of the target service in the Completed status. The service details page is displayed.
- You can view the service name, status, ID, input path, output path, and description.
- You can click
in the Description area to edit the description.
+ - Obtain the detailed OBS path next to Output Path, switch to the path and obtain the batch service prediction results, including the prediction result file and the AI application prediction result.
If the prediction is successful, the directory contains the prediction result file and AI application prediction result. Otherwise, the directory contains only the prediction result file.
+- Prediction result file: The file is in xxx.manifest format, which contains the file path and prediction result, and more.
- AI application prediction result:
- If images are input, a result file is generated for each image in the Image name__result.txt format, for example, IMG_20180919_115016.jpg_result.txt.
- If audio files are input, a result file is generated for each audio file in the Audio file name__result.txt format, for example, 1-36929-A-47.wav_result.txt.
- If table data is input, the result file is generated in the Table name__result.txt format, for example, train.csv_result.txt.
+
+
+
+
+
+
diff --git a/docs/modelarts/umn/inference-modelarts-0053.html b/docs/modelarts/umn/inference-modelarts-0053.html
new file mode 100644
index 00000000..14373405
--- /dev/null
+++ b/docs/modelarts/umn/inference-modelarts-0053.html
@@ -0,0 +1,21 @@
+
+
+
+
Inference Specifications
+
+
+
+
+
diff --git a/docs/modelarts/umn/inference-modelarts-0054.html b/docs/modelarts/umn/inference-modelarts-0054.html
new file mode 100644
index 00000000..dcb9bdbf
--- /dev/null
+++ b/docs/modelarts/umn/inference-modelarts-0054.html
@@ -0,0 +1,20 @@
+
+
+
Model Package Specifications
+
+
+
diff --git a/docs/modelarts/umn/inference-modelarts-0055.html b/docs/modelarts/umn/inference-modelarts-0055.html
new file mode 100644
index 00000000..e16f65c4
--- /dev/null
+++ b/docs/modelarts/umn/inference-modelarts-0055.html
@@ -0,0 +1,52 @@
+
+
+
Introduction
+
When creating an AI application on the AI application management page, make sure that any meta model imported from OBS complies with certain specifications.
+
+
The model package must contain the model directory. The model directory stores the model file, model configuration file, and model inference code file.
+
- Model files: The requirements for model files vary according to the model package structure. For details, see Model Package Example.
- Model configuration file: The model configuration file must be available and its name is consistently to be config.json. There must be only one model configuration file. For details about how to edit a model configuration file, see Specifications for Editing a Model Configuration File.
- Model inference code file: It is mandatory. The file name is consistently to be customize_service.py. There must be only one model inference code file. For details about how to edit model inference code, see Specifications for Writing Model Inference Code.
- The .py file on which customize_service.py depends can be directly stored in the model directory. Use relative import for the custom package.
- The other files on which customize_service.py depends can be stored in the model directory. Only absolute paths can be used to access these files. For details, see Obtaining an Absolute Path.
+
+
ModelArts also provides custom script examples of common AI engines. For details, see Examples of Custom Scripts.
+
Model Package Example
- Structure of the TensorFlow-based model package
When publishing the model, you only need to specify the ocr directory.
+OBS bucket/directory name
+|── ocr
+| ├── model (Mandatory) Name of a fixed subdirectory, which is used to store model-related files
+| │ ├── <<Custom Python package>> (Optional) Your Python package, which can be directly referenced in model inference code
+| │ ├── saved_model.pb (Mandatory) Protocol buffer file, which contains the diagram description of the model
+| │ ├── variables Name of a fixed sub-directory, which contains the weight and deviation rate of the model. It is mandatory for the main file of a *.pb model.
+| │ │ ├── variables.index Mandatory
+| │ │ ├── variables.data-00000-of-00001 Mandatory
+| │ ├──config.json (Mandatory) Model configuration file. The file name is fixed to config.json. Only one model configuration file is allowed.
+| │ ├──customize_service.py (Mandatory) Model inference code. The file name is fixed to customize_service.py. Only one model inference code file exists. The files on which customize_service.py depends can be directly stored in the model directory.
+ - Structure of the MindSpore-based model package
OBS bucket/directory name
+|── resnet
+| ├── model (Mandatory) Name of a fixed subdirectory, which is used to store model-related files
+| │ ├── <<Custom Python package>> (Optional) Your Python package, which can be directly referenced in model inference code
+| │ ├── checkpoint_lenet_1-1_1875.ckpt (Mandatory) Model file in ckpt format trained using MindSpore
+| │ ├── config.json (Mandatory) Model configuration file. The file name is fixed to config.json. Only one model configuration file is allowed.
+| │ ├── customize_service.py (Mandatory) Model inference code. The file name is fixed to customize_service.py. Only one model inference code file exists. The files on which customize_service.py depends can be directly stored in the model directory.
+| │ ├── tmp.om (Mandatory) An empty .om file that enables the model package to be imported
+
+ - Structure of the image-based model package
When publishing the model, you only need to specify the resnet directory.
+OBS bucket/directory name
+|── resnet
+| ├── model (Mandatory) Name of a fixed subdirectory, which is used to store model-related files
+| │ ├──config.json (Mandatory) Model configuration file (the address of the SWR image must be configured). The file name is fixed to config.json. Only one model configuration file is allowed.
+ - Structure of the PyTorch-based model package
When publishing the model, you only need to specify the resnet directory.
+OBS bucket/directory name
+|── resnet
+| ├── model (Mandatory) Name of a fixed subdirectory, which is used to store model-related files
+| │ ├── <<Custom Python package>> (Optional) Your Python package, which can be directly referenced in model inference code
+| │ ├── resnet50.pth (Mandatory) PyTorch model file, which contains variable and weight information and is saved as state_dict
+| │ ├──config.json (Mandatory) Model configuration file. The file name is fixed to config.json. Only one model configuration file is allowed.
+| │ ├──customize_service.py (Mandatory) Model inference code. The file name is fixed to customize_service.py. Only one model inference code file exists. The files on which customize_service.py depends can be directly stored in the model directory.
+
+
+
+
+
diff --git a/docs/modelarts/umn/inference-modelarts-0056.html b/docs/modelarts/umn/inference-modelarts-0056.html
new file mode 100644
index 00000000..30229955
--- /dev/null
+++ b/docs/modelarts/umn/inference-modelarts-0056.html
@@ -0,0 +1,1406 @@
+
+
+
Specifications for Editing a Model Configuration File
+
You must edit a configuration file config.json when publishing a model. The model configuration file describes the model usage, computing framework, precision, inference code dependency package, and model API.
+
Configuration File Format
The configuration file is in JSON format. Table 1 describes the parameters.
+
+
Table 1 ParametersParameter
+ |
+Mandatory
+ |
+Type
+ |
+Description
+ |
+
|---|
+
+model_algorithm
+ |
+Yes
+ |
+String
+ |
+Model algorithm, which is set by the model developer to help users understand the usage of the model. The value must start with a letter and contain no more than 36 characters. Special characters (&!'\"<>=) are not allowed. Common model algorithms include image_classification, object_detection, and predict_analysis.
+ |
+
+model_type
+ |
+Yes
+ |
+String
+ |
+Model AI engine, which indicates the computing framework used by a model. Common AI engines and Image are supported.
+
+ |
+
+runtime
+ |
+No
+ |
+String
+ |
+Model runtime environment. Python 3.6 is used by default. The value of runtime depends on the value of model_type. If model_type is set to Image, you do not need to set runtime. If model_type is set to another mainstream framework, select the engine and runtime environment.
+ |
+
+swr_location
+ |
+No
+ |
+String
+ |
+SWR image address.
+- If you import a custom image metamodel from a container image, you do not need to set swr_location.
- If you import a custom image meta model from OBS (not recommended) and set model_type to Image, you must set swr_location. swr_location specifies the path to the Docker image on SWR, which will be used to publish the model.
+ |
+
+metrics
+ |
+No
+ |
+Object
+ |
+Model precision information, including the F1 score, recall, precision, and accuracy. For details about the metrics object structure, see Table 2.
+The result is displayed in the model precision area on the AI application details page.
+ |
+
+apis
+ |
+No
+ |
+api array
+ |
+Format of the requests received and returned by a model. The value is structure data.
+It is the RESTful API array provided by a model. For details about the API data structure, see Table 3. For details about the code example, see Code Example of apis Parameters.
+- If model_type is set to Image, the AI application is created using a custom image. APIs with different paths can be declared in apis based on the request path exposed by the image.
- When model_type is not Image, only one API whose request path is / can be declared in apis because the preconfigured AI engine exposes only one inference API whose request path is /.
+ |
+
+dependencies
+ |
+No
+ |
+dependency array
+ |
+Package on which the model inference code depends, which is structure data.
+Model developers need to provide the package name, installation mode, and version constraints. Only the pip installation mode is supported. Table 6 describes the dependency array.
+If the model package does not contain the customize_service.py file, you do not need to set this parameter. Dependency packages cannot be installed for custom image models.
+ |
+
+health
+ |
+No
+ |
+health data structure
+ |
+Configuration of an image health interface. This parameter is mandatory only when model_type is set to Image.
+If services cannot be interrupted during a rolling upgrade, a health check API must be provided for ModelArts to call. For details about the health data structure, see Table 8.
+ |
+
+
+
+
+
+
Table 2 metrics object descriptionParameter
+ |
+Mandatory
+ |
+Type
+ |
+Description
+ |
+
|---|
+
+f1
+ |
+No
+ |
+Number
+ |
+F1 score. The value is rounded to 17 decimal places.
+ |
+
+recall
+ |
+No
+ |
+Number
+ |
+Recall. The value is rounded to 17 decimal places.
+ |
+
+precision
+ |
+No
+ |
+Number
+ |
+Precision. The value is rounded to 17 decimal places.
+ |
+
+accuracy
+ |
+No
+ |
+Number
+ |
+Accuracy. The value is rounded to 17 decimal places.
+ |
+
+
+
+
+
+
Table 3 api data structure descriptionParameter
+ |
+Mandatory
+ |
+Type
+ |
+Description
+ |
+
|---|
+
+url
+ |
+No
+ |
+String
+ |
+Request path. The default value is a slash (/). For a custom image model (model_type is Image), set this parameter to the actual request path exposed in the image. For a non-custom image model (model_type is not Image), the URL can only be /.
+ |
+
+method
+ |
+No
+ |
+String
+ |
+Request method. The default value is POST.
+ |
+
+request
+ |
+No
+ |
+Object
+ |
+Request body. For details, see Table 4.
+ |
+
+response
+ |
+No
+ |
+Object
+ |
+Response body. For details, see Table 5.
+ |
+
+
+
+
+
+
Table 4 request structure descriptionParameter
+ |
+Mandatory
+ |
+Type
+ |
+Description
+ |
+
|---|
+
+Content-type
+ |
+No for real-time services
+Yes for batch services
+ |
+String
+ |
+Data is sent in a specified content format. The default value is application/json.
+The options are as follows:
+- application/json: JSON data is uploaded.
- multipart/form-data: A file is uploaded.
+ NOTE: For machine learning models, only application/json is supported.
+
+ |
+
+data
+ |
+No for real-time services
+Yes for batch services
+ |
+String
+ |
+The request body is described in JSON schema.
+ |
+
+
+
+
+
+
Table 5 response structure descriptionParameter
+ |
+Mandatory
+ |
+Type
+ |
+Description
+ |
+
|---|
+
+Content-type
+ |
+No for real-time services
+Yes for batch services
+ |
+String
+ |
+Data is sent in a specified content format. The default value is application/json.
+ NOTE: For machine learning models, only application/json is supported.
+
+ |
+
+data
+ |
+No for real-time services
+Yes for batch services
+ |
+String
+ |
+The response body is described in JSON schema.
+ |
+
+
+
+
+
+
Table 6 dependency array descriptionParameter
+ |
+Mandatory
+ |
+Type
+ |
+Description
+ |
+
|---|
+
+installer
+ |
+Yes
+ |
+String
+ |
+Installation method. Only pip is supported.
+ |
+
+packages
+ |
+Yes
+ |
+package array
+ |
+Dependency package collection. For details about the package structure array, see Table 7.
+ |
+
+
+
+
+
+
Table 7 package array descriptionParameter
+ |
+Mandatory
+ |
+Type
+ |
+Description
+ |
+
|---|
+
+package_name
+ |
+Yes
+ |
+String
+ |
+Dependency package name. Special characters (&!'"<>=) are not allowed.
+ |
+
+package_version
+ |
+No
+ |
+String
+ |
+Dependency package version. If the dependency package does not rely on package versions, leave this field blank. Special characters (&!'"<>=) are not allowed.
+ |
+
+restraint
+ |
+No
+ |
+String
+ |
+Version restriction. This parameter is mandatory only when package_version is configured. Possible values are EXACT, ATLEAST, and ATMOST.
+- EXACT indicates that a specified version is installed.
- ATLEAST indicates that the version of the installation package is not earlier than the specified version.
- ATMOST indicates that the version of the installation package is not later than the specified version.
NOTE: - If there are specific requirements on the version, preferentially use EXACT. If EXACT conflicts with the system installation packages, you can select ATLEAST.
- If there is no specific requirement on the version, retain only the package_name parameter and leave restraint and package_version blank.
+
+
+ |
+
+
+
+
+
+
Table 8 health data structure descriptionParameter
+ |
+Mandatory
+ |
+Type
+ |
+Description
+ |
+
|---|
+
+url
+ |
+Yes
+ |
+String
+ |
+Request URL of the health check API.
+ |
+
+protocol
+ |
+No
+ |
+String
+ |
+Request protocol of the health check API. Only HTTP is supported.
+ |
+
+initial_delay_seconds
+ |
+No
+ |
+String
+ |
+After an instance is started, a health check starts after seconds configured in initial_delay_seconds.
+ |
+
+timeout_seconds
+ |
+No
+ |
+String
+ |
+Health check timeout duration in the unit of second. This parameter cannot be left blank.
+ |
+
+
+
+
+
+
Code Example of apis Parameters
[{
+ "url": "/",
+ "method": "post",
+ "request": {
+ "Content-type": "multipart/form-data",
+ "data": {
+ "type": "object",
+ "properties": {
+ "images": {
+ "type": "file"
+ }
+ }
+ }
+ },
+ "response": {
+ "Content-type": "applicaton/json",
+ "data": {
+ "type": "object",
+ "properties": {
+ "mnist_result": {
+ "type": "array",
+ "item": [
+ {
+ "type": "string"
+ }
+ ]
+ }
+ }
+ }
+ }
+}]
+
+
Example of an Object Detection Model Configuration File
The following code uses the TensorFlow engine as an example. You can modify the model_type parameter based on the actual engine type.
+
- Model input
Key: images
+Value: image files
+ - Model output
1
+ 2
+ 3
+ 4
+ 5
+ 6
+ 7
+ 8
+ 9
+10
+11
+12
+13
+14
+15
+16
+17
+18
+19
+20
+21 | {
+ "detection_classes": [
+ "face",
+ "arm"
+ ],
+ "detection_boxes": [
+ [
+ 33.6,
+ 42.6,
+ 104.5,
+ 203.4
+ ],
+ [
+ 103.1,
+ 92.8,
+ 765.6,
+ 945.7
+ ]
+ ],
+ "detection_scores": [0.99, 0.73]
+}
+ |
+
+
- Configuration file
1
+ 2
+ 3
+ 4
+ 5
+ 6
+ 7
+ 8
+ 9
+10
+11
+12
+13
+14
+15
+16
+17
+18
+19
+20
+21
+22
+23
+24
+25
+26
+27
+28
+29
+30
+31
+32
+33
+34
+35
+36
+37
+38
+39
+40
+41
+42
+43
+44
+45
+46
+47
+48
+49
+50
+51
+52
+53
+54
+55
+56
+57
+58
+59
+60
+61
+62
+63
+64
+65
+66
+67
+68
+69
+70
+71 | {
+ "model_type": "TensorFlow",
+ "model_algorithm": "object_detection",
+ "metrics": {
+ "f1": 0.345294,
+ "accuracy": 0.462963,
+ "precision": 0.338977,
+ "recall": 0.351852
+ },
+ "apis": [{
+
+ "url": "/",
+ "method": "post",
+ "request": {
+ "Content-type": "multipart/form-data",
+ "data": {
+ "type": "object",
+ "properties": {
+ "images": {
+ "type": "file"
+ }
+ }
+ }
+ },
+ "response": {
+ "Content-type": "application/json",
+ "data": {
+ "type": "object",
+ "properties": {
+ "detection_classes": {
+ "type": "array",
+ "items": [{
+ "type": "string"
+ }]
+ },
+ "detection_boxes": {
+ "type": "array",
+ "items": [{
+ "type": "array",
+ "minItems": 4,
+ "maxItems": 4,
+ "items": [{
+ "type": "number"
+ }]
+ }]
+ },
+ "detection_scores": {
+ "type": "array",
+ "items": [{
+ "type": "number"
+ }]
+ }
+ }
+ }
+ }
+ }],
+ "dependencies": [{
+ "installer": "pip",
+ "packages": [{
+ "restraint": "EXACT",
+ "package_version": "1.15.0",
+ "package_name": "numpy"
+ },
+ {
+ "restraint": "EXACT",
+ "package_version": "5.2.0",
+ "package_name": "Pillow"
+ }
+ ]
+ }]
+}
+ |
+
+
+
+
Example of an Image Classification Model Configuration File
The following code uses the TensorFlow engine as an example. You can modify the model_type parameter based on the actual engine type.
+
- Model input
Key: images
+Value: image files
+ - Model output
| {
+ "predicted_label": "flower",
+ "scores": [
+ ["rose", 0.99],
+ ["begonia", 0.01]
+ ]
+}
+ |
+
+
- Configuration file
1
+ 2
+ 3
+ 4
+ 5
+ 6
+ 7
+ 8
+ 9
+10
+11
+12
+13
+14
+15
+16
+17
+18
+19
+20
+21
+22
+23
+24
+25
+26
+27
+28
+29
+30
+31
+32
+33
+34
+35
+36
+37
+38
+39
+40
+41
+42
+43
+44
+45
+46
+47
+48
+49
+50
+51
+52
+53
+54
+55
+56
+57
+58
+59
+60
+61
+62
+63
+64
+65
+66
+67 | {
+ "model_type": "TensorFlow",
+ "model_algorithm": "image_classification",
+ "metrics": {
+ "f1": 0.345294,
+ "accuracy": 0.462963,
+ "precision": 0.338977,
+ "recall": 0.351852
+ },
+ "apis": [{
+
+ "url": "/",
+ "method": "post",
+ "request": {
+ "Content-type": "multipart/form-data",
+ "data": {
+ "type": "object",
+ "properties": {
+ "images": {
+ "type": "file"
+ }
+ }
+ }
+ },
+ "response": {
+ "Content-type": "application/json",
+ "data": {
+ "type": "object",
+ "properties": {
+ "predicted_label": {
+ "type": "string"
+ },
+ "scores": {
+ "type": "array",
+ "items": [{
+ "type": "array",
+ "minItems": 2,
+ "maxItems": 2,
+ "items": [
+ {
+ "type": "string"
+ },
+ {
+ "type": "number"
+ }
+ ]
+ }]
+ }
+ }
+ }
+ }
+ }],
+ "dependencies": [{
+ "installer": "pip",
+ "packages": [{
+ "restraint": "ATLEAST",
+ "package_version": "1.15.0",
+ "package_name": "numpy"
+ },
+ {
+ "restraint": "",
+ "package_version": "",
+ "package_name": "Pillow"
+ }
+ ]
+ }]
+}
+ |
+
+
+
+
Example of a Predictive Analytics Model Configuration File
The following code uses the TensorFlow engine as an example. You can modify the model_type parameter based on the actual engine type.
+
- Model input
1
+ 2
+ 3
+ 4
+ 5
+ 6
+ 7
+ 8
+ 9
+10
+11
+12
+13
+14
+15
+16
+17
+18
+19
+20
+21
+22
+23
+24 | {
+ "data": {
+ "req_data": [
+ {
+ "buying_price": "high",
+ "maint_price": "high",
+ "doors": "2",
+ "persons": "2",
+ "lug_boot": "small",
+ "safety": "low",
+ "acceptability": "acc"
+ },
+ {
+ "buying_price": "high",
+ "maint_price": "high",
+ "doors": "2",
+ "persons": "2",
+ "lug_boot": "small",
+ "safety": "low",
+ "acceptability": "acc"
+ }
+ ]
+ }
+}
+ |
+
+
- Model output
1
+ 2
+ 3
+ 4
+ 5
+ 6
+ 7
+ 8
+ 9
+10
+11
+12 | {
+ "data": {
+ "resp_data": [
+ {
+ "predict_result": "unacc"
+ },
+ {
+ "predict_result": "unacc"
+ }
+ ]
+ }
+}
+ |
+
+
- Configuration file
In the code, the data parameter in the request and response structures is described in JSON Schema. The content in data and properties corresponds to the model input and output.
+
+ 1
+ 2
+ 3
+ 4
+ 5
+ 6
+ 7
+ 8
+ 9
+10
+11
+12
+13
+14
+15
+16
+17
+18
+19
+20
+21
+22
+23
+24
+25
+26
+27
+28
+29
+30
+31
+32
+33
+34
+35
+36
+37
+38
+39
+40
+41
+42
+43
+44
+45
+46
+47
+48
+49
+50
+51
+52
+53
+54
+55
+56
+57
+58
+59
+60
+61
+62
+63
+64
+65
+66
+67
+68
+69
+70
+71
+72
+73
+74
+75
+76
+77
+78 | {
+ "model_type": "TensorFlow",
+ "model_algorithm": "predict_analysis",
+ "metrics": {
+ "f1": 0.345294,
+ "accuracy": 0.462963,
+ "precision": 0.338977,
+ "recall": 0.351852
+ },
+ "apis": [
+ {
+
+ "url": "/",
+ "method": "post",
+ "request": {
+ "Content-type": "application/json",
+ "data": {
+ "type": "object",
+ "properties": {
+ "data": {
+ "type": "object",
+ "properties": {
+ "req_data": {
+ "items": [
+ {
+ "type": "object",
+ "properties": {}
+ }
+ ],
+ "type": "array"
+ }
+ }
+ }
+ }
+ }
+ },
+ "response": {
+ "Content-type": "application/json",
+ "data": {
+ "type": "object",
+ "properties": {
+ "data": {
+ "type": "object",
+ "properties": {
+ "resp_data": {
+ "type": "array",
+ "items": [
+ {
+ "type": "object",
+ "properties": {}
+ }
+ ]
+ }
+ }
+ }
+ }
+ }
+ }
+ }
+ ],
+ "dependencies": [
+ {
+ "installer": "pip",
+ "packages": [
+ {
+ "restraint": "EXACT",
+ "package_version": "1.15.0",
+ "package_name": "numpy"
+ },
+ {
+ "restraint": "EXACT",
+ "package_version": "5.2.0",
+ "package_name": "Pillow"
+ }
+ ]
+ }
+ ]
+}
+ |
+
+
+
+
Example of a Custom Image Model Configuration File
The model input and output are similar to those in Example of an Object Detection Model Configuration File.
+
- The following is a request example when the input is images.
This is a model prediction request containing the parameter images with the parameter type of file is received. The file upload button is displayed on the inference page. You need to upload files for inference.
+ 1
+ 2
+ 3
+ 4
+ 5
+ 6
+ 7
+ 8
+ 9
+10
+11 | {
+ "Content-type": "multipart/form-data",
+ "data": {
+ "type": "object",
+ "properties": {
+ "images": {
+ "type": "file"
+ }
+ }
+ }
+}
+ |
+
+
- The following is a request example when the input is JSON data.
This is a JSON request body for model prediction. There is only one prediction request containing the parameter input with the parameter type of string. On the inference page, a text box is displayed for you to enter the prediction request.
+ 1
+ 2
+ 3
+ 4
+ 5
+ 6
+ 7
+ 8
+ 9
+10
+11 | {
+ "Content-type": "application/json",
+ "data": {
+ "type": "object",
+ "properties": {
+ "input": {
+ "type": "string"
+ }
+ }
+ }
+}
+ |
+
+
+
A complete request example is as follows:
+
1
+ 2
+ 3
+ 4
+ 5
+ 6
+ 7
+ 8
+ 9
+10
+11
+12
+13
+14
+15
+16
+17
+18
+19
+20
+21
+22
+23
+24
+25
+26
+27
+28
+29
+30
+31
+32
+33
+34
+35
+36
+37
+38
+39
+40
+41
+42
+43
+44
+45
+46
+47
+48
+49
+50
+51
+52
+53
+54
+55
+56 | {
+ "model_algorithm": "image_classification",
+ "model_type": "Image",
+ "metrics": {
+ "f1": 0.345294,
+ "accuracy": 0.462963,
+ "precision": 0.338977,
+ "recall": 0.351852
+ },
+ "apis": [{
+
+ "url": "/",
+ "method": "post",
+ "request": {
+ "Content-type": "multipart/form-data",
+ "data": {
+ "type": "object",
+ "properties": {
+ "images": {
+ "type": "file"
+ }
+ }
+ }
+ },
+ "response": {
+ "Content-type": "application/json",
+ "data": {
+ "type": "object",
+ "required": [
+ "predicted_label",
+ "scores"
+ ],
+ "properties": {
+ "predicted_label": {
+ "type": "string"
+ },
+ "scores": {
+ "type": "array",
+ "items": [{
+ "type": "array",
+ "minItems": 2,
+ "maxItems": 2,
+ "items": [{
+ "type": "string"
+ },
+ {
+ "type": "number"
+ }
+ ]
+ }]
+ }
+ }
+ }
+ }
+ }]
+}
+ |
+
+
+
+
Example of a Machine Learning Model Configuration File
The following uses XGBoost as an example:
+
+
{
+ "req_data": [
+ {
+ "sepal_length": 5,
+ "sepal_width": 3.3,
+ "petal_length": 1.4,
+ "petal_width": 0.2
+ },
+ {
+ "sepal_length": 5,
+ "sepal_width": 2,
+ "petal_length": 3.5,
+ "petal_width": 1
+ },
+ {
+ "sepal_length": 6,
+ "sepal_width": 2.2,
+ "petal_length": 5,
+ "petal_width": 1.5
+ }
+ ]
+}
+
+
{
+ "resp_data": [
+ {
+ "predict_result": "Iris-setosa"
+ },
+ {
+ "predict_result": "Iris-versicolor"
+ }
+ ]
+}
+
+
{
+ "model_type": "XGBoost",
+ "model_algorithm": "xgboost_iris_test",
+ "runtime": "python2.7",
+ "metrics": {
+ "f1": 0.345294,
+ "accuracy": 0.462963,
+ "precision": 0.338977,
+ "recall": 0.351852
+ },
+ "apis": [
+ {
+
+ "url": "/",
+ "method": "post",
+ "request": {
+ "Content-type": "application/json",
+ "data": {
+ "type": "object",
+ "properties": {
+ "req_data": {
+ "items": [
+ {
+ "type": "object",
+ "properties": {}
+ }
+ ],
+ "type": "array"
+ }
+ }
+ }
+ },
+ "response": {
+ "Content-type": "applicaton/json",
+ "data": {
+ "type": "object",
+ "properties": {
+ "resp_data": {
+ "type": "array",
+ "items": [
+ {
+ "type": "object",
+ "properties": {
+ "predict_result": {}
+ }
+ }
+ ]
+ }
+ }
+ }
+ }
+ }
+ ]
+}
+
+
+
Example of a Model Configuration File Using a Custom Dependency Package
The following example defines the NumPy 1.16.4 dependency environment.
+
1
+ 2
+ 3
+ 4
+ 5
+ 6
+ 7
+ 8
+ 9
+10
+11
+12
+13
+14
+15
+16
+17
+18
+19
+20
+21
+22
+23
+24
+25
+26
+27
+28
+29
+30
+31
+32
+33
+34
+35
+36
+37
+38
+39
+40
+41
+42
+43
+44
+45
+46
+47
+48
+49
+50
+51
+52
+53
+54
+55
+56
+57 | {
+ "model_algorithm": "image_classification",
+ "model_type": "TensorFlow",
+ "runtime": "python3.6",
+ "apis": [
+ {
+ "procotol": "http",
+ "url": "/",
+ "method": "post",
+ "request": {
+ "Content-type": "multipart/form-data",
+ "data": {
+ "type": "object",
+ "properties": {
+ "images": {
+ "type": "file"
+ }
+ }
+ }
+ },
+ "response": {
+ "Content-type": "applicaton/json",
+ "data": {
+ "type": "object",
+ "properties": {
+ "mnist_result": {
+ "type": "array",
+ "item": [
+ {
+ "type": "string"
+ }
+ ]
+ }
+ }
+ }
+ }
+ }
+ ],
+ "metrics": {
+ "f1": 0.124555,
+ "recall": 0.171875,
+ "precision": 0.00234938928519385,
+ "accuracy": 0.00746268656716417
+ },
+ "dependencies": [
+ {
+ "installer": "pip",
+ "packages": [
+ {
+ "restraint": "EXACT",
+ "package_version": "1.16.4",
+ "package_name": "numpy"
+ }
+ ]
+ }
+ ]
+}
+ |
+
+
+
+
+
+
diff --git a/docs/modelarts/umn/inference-modelarts-0057.html b/docs/modelarts/umn/inference-modelarts-0057.html
new file mode 100644
index 00000000..2a871923
--- /dev/null
+++ b/docs/modelarts/umn/inference-modelarts-0057.html
@@ -0,0 +1,537 @@
+
+
+
Specifications for Writing Model Inference Code
+
This section describes the general method of editing model inference code in ModelArts. This section also provides an inference code example for the TensorFlow engine and an example of customizing the inference logic in the inference script.
+
Due to the limitation of API Gateway, the duration of a single prediction in ModelArts cannot exceed 40s. The model inference code must be logically clear and concise for satisfactory inference performance.
+
Specifications for Writing Inference Code
- In the model inference code file customize_service.py, add a child model class. This child model class inherits properties from its parent model class. For details about the import statements of different types of parent model classes, see Table 1.
+
Table 1 Import statements of different types of parent model classesModel Type
+ |
+Parent Class
+ |
+Import Statement
+ |
+
|---|
+
+TensorFlow
+ |
+TfServingBaseService
+ |
+from model_service.tfserving_model_service import TfServingBaseService
+ |
+
+PyTorch
+ |
+PTServingBaseService
+ |
+from model_service.pytorch_model_service import PTServingBaseService
+ |
+
+MindSpore
+ |
+SingleNodeService
+ |
+from model_service.model_service import SingleNodeService
+ |
+
+
+
+
- The following methods can be overridden.
+
Table 2 Methods to be overriddenMethod
+ |
+Description
+ |
+
|---|
+
+__init__(self, model_name, model_path)
+ |
+Initialization method, which is suitable for models created based on deep learning frameworks. Models and labels are loaded using this method. This method must be overridden for models based on PyTorch and Caffe to implement the model loading logic.
+ |
+
+__init__(self, model_path)
+ |
+Initialization method, which is suitable for models created based on machine learning frameworks. This method initializes the model path (self.model_path). In Spark_MLlib, this method also initializes SparkSession (self.spark).
+ |
+
+_preprocess(self, data)
+ |
+Preprocess method, which is called before an inference request and is used to convert the original request data of an API into the expected input data of a model.
+ |
+
+_inference(self, data)
+ |
+Inference request method. You are advised not to override the method because once the method is overridden, the built-in inference process of ModelArts will be overwritten and the custom inference logic will run.
+ |
+
+_postprocess(self, data)
+ |
+Postprocess method, which is called after an inference request is complete and is used to convert the model output to the API output.
+ |
+
+
+
+
- You can override the preprocess and postprocess methods for preprocessing the API input and postprocessing the inference output.
- Overriding the init method of the parent model class may cause an AI application to run abnormally.
+
+ - The attribute that can be used is the local path where the model resides. The attribute name is self.model_path. In addition, PySpark-based models can use self.spark to obtain the SparkSession object in customize_service.py.
An absolute path is required for reading files in the inference code. You can obtain the local path of the model from the self.model_path attribute.
+
+
+
+ - data imported through the API for pre-processing, actual inference request, and post-processing can be multipart/form-data or application/json.
- multipart/form-data request
curl -X POST \
+ <modelarts-inference-endpoint> \
+ -F image1=@cat.jpg \
+ -F images2=@horse.jpg
+The input data is as follows:
+[
+ {
+ "image1":{
+ "cat.jpg":"<cat.jpg file io>"
+ }
+ },
+ {
+ "image2":{
+ "horse.jpg":"<horse.jpg file io>"
+ }
+ }
+]
+ - application/json request
curl -X POST \
+ <modelarts-inference-endpoint> \
+ -d '{
+ "images":"base64 encode image"
+ }'
+The input data is python dict.
+ {
+ "images":"base64 encode image"
+
+ }
+
+
+
+
TensorFlow Inference Script Example
The following is an example of TensorFlow MnistService. For details about the inference code of other engines, see
PyTorch and
Caffe.
+
+
The preceding sample code resizes images imported to the user's form to adapt to the model input shape. The 32×32 image is read from the Pillow library and resized to 1×784 to match the model input. In subsequent processing, convert the model output into a list for the RESTful API to display.
+
+
Inference Script Example of Custom Inference Logic
Customize a dependency package in the configuration file by referring to Example of a Model Configuration File Using a Custom Dependency Package. Then, use the following code example to load the model in saved_model format for inference.
+
Python logging used by base inference images allows the display of only warning logs. To display INFO logs, set the log level to INFO in the code.
+
+
1
+ 2
+ 3
+ 4
+ 5
+ 6
+ 7
+ 8
+ 9
+ 10
+ 11
+ 12
+ 13
+ 14
+ 15
+ 16
+ 17
+ 18
+ 19
+ 20
+ 21
+ 22
+ 23
+ 24
+ 25
+ 26
+ 27
+ 28
+ 29
+ 30
+ 31
+ 32
+ 33
+ 34
+ 35
+ 36
+ 37
+ 38
+ 39
+ 40
+ 41
+ 42
+ 43
+ 44
+ 45
+ 46
+ 47
+ 48
+ 49
+ 50
+ 51
+ 52
+ 53
+ 54
+ 55
+ 56
+ 57
+ 58
+ 59
+ 60
+ 61
+ 62
+ 63
+ 64
+ 65
+ 66
+ 67
+ 68
+ 69
+ 70
+ 71
+ 72
+ 73
+ 74
+ 75
+ 76
+ 77
+ 78
+ 79
+ 80
+ 81
+ 82
+ 83
+ 84
+ 85
+ 86
+ 87
+ 88
+ 89
+ 90
+ 91
+ 92
+ 93
+ 94
+ 95
+ 96
+ 97
+ 98
+ 99
+100
+101
+102
+103
+104
+105
+106
+107
+108
+109
+110
+111
+112
+113
+114 | # -*- coding: utf-8 -*-
+import json
+import os
+import threading
+
+import numpy as np
+import tensorflow as tf
+from PIL import Image
+
+from model_service.tfserving_model_service import TfServingBaseService
+import logging
+logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(name)s - %(levelname)s - %(message)s')
+
+logger = logging.getLogger(__name__)
+
+
+class MnistService(TfServingBaseService):
+
+ def __init__(self, model_name, model_path):
+ self.model_name = model_name
+ self.model_path = model_path
+ self.model_inputs = {}
+ self.model_outputs = {}
+
+ # The label file can be loaded here and used in the post-processing function.
+ # Directories for storing the label.txt file on OBS and in the model package
+
+ # with open(os.path.join(self.model_path, 'label.txt')) as f:
+ # self.label = json.load(f)
+
+ # Load the model in saved_model format in non-blocking mode to prevent blocking timeout.
+ thread = threading.Thread(target=self.get_tf_sess)
+ thread.start()
+
+ def get_tf_sess(self):
+ # Load the model in saved_model format.
+
+ # The session will be reused. Do not use the with statement.
+ sess = tf.Session(graph=tf.Graph())
+ meta_graph_def = tf.saved_model.loader.load(sess, [tf.saved_model.tag_constants.SERVING], self.model_path)
+ signature_defs = meta_graph_def.signature_def
+
+ self.sess = sess
+
+ signature = []
+
+ # only one signature allowed
+ for signature_def in signature_defs:
+ signature.append(signature_def)
+ if len(signature) == 1:
+ model_signature = signature[0]
+ else:
+ logger.warning("signatures more than one, use serving_default signature")
+ model_signature = tf.saved_model.signature_constants.DEFAULT_SERVING_SIGNATURE_DEF_KEY
+
+ logger.info("model signature: %s", model_signature)
+
+ for signature_name in meta_graph_def.signature_def[model_signature].inputs:
+ tensorinfo = meta_graph_def.signature_def[model_signature].inputs[signature_name]
+ name = tensorinfo.name
+ op = self.sess.graph.get_tensor_by_name(name)
+ self.model_inputs[signature_name] = op
+
+ logger.info("model inputs: %s", self.model_inputs)
+
+ for signature_name in meta_graph_def.signature_def[model_signature].outputs:
+ tensorinfo = meta_graph_def.signature_def[model_signature].outputs[signature_name]
+ name = tensorinfo.name
+ op = self.sess.graph.get_tensor_by_name(name)
+
+ self.model_outputs[signature_name] = op
+
+ logger.info("model outputs: %s", self.model_outputs)
+
+ def _preprocess(self, data):
+ # Two request modes using HTTPS
+ # 1. The request in form-data format is as follows: data = {"Request key value":{"File name":<File io>}}
+ # 2. Request in JSON format is as follows: data = json.loads("JSON body passed in the API")
+ preprocessed_data = {}
+
+ for k, v in data.items():
+ for file_name, file_content in v.items():
+ image1 = Image.open(file_content)
+ image1 = np.array(image1, dtype=np.float32)
+ image1.resize((1, 28, 28))
+ preprocessed_data[k] = image1
+
+ return preprocessed_data
+
+ def _inference(self, data):
+
+ feed_dict = {}
+ for k, v in data.items():
+ if k not in self.model_inputs.keys():
+ logger.error("input key %s is not in model inputs %s", k, list(self.model_inputs.keys()))
+ raise Exception("input key %s is not in model inputs %s" % (k, list(self.model_inputs.keys())))
+ feed_dict[self.model_inputs[k]] = v
+
+ result = self.sess.run(self.model_outputs, feed_dict=feed_dict)
+ logger.info('predict result : ' + str(result))
+
+ return result
+
+ def _postprocess(self, data):
+ infer_output = {"mnist_result": []}
+ for output_name, results in data.items():
+
+ for result in results:
+ infer_output["mnist_result"].append(np.argmax(result))
+
+ return infer_output
+
+ def __del__(self):
+ self.sess.close()
+ |
+
+
+
+
To load models that are not supported by ModelArts or multiple models, specify the loading path using the __init__ method. Example code is as follows:
+
# -*- coding: utf-8 -*-
+import os
+from model_service.tfserving_model_service import TfServingBaseService
+
+class MnistService(TfServingBaseService):
+ def __init__(self, model_name, model_path):
+ # Obtain the path to the model folder.
+ root = os.path.dirname(os.path.abspath(__file__))
+ # test.onnx is the name of the model file to be loaded and must be stored in the model folder.
+ self.model_path = os.path.join(root, test.onnx)
+
+ # Loading multiple models, for example, test2.onnx
+ # self.model_path2 = os.path.join(root, test2.onnx)
+
+
MindSpore Inference Script Example
The inference script is as follows:
import threading
+
+import mindspore
+import mindspore.nn as nn
+import numpy as np
+import logging
+from mindspore import Tensor, context
+from mindspore.common.initializer import Normal
+from mindspore.train.serialization import load_checkpoint, load_param_into_net
+from model_service.model_service import SingleNodeService
+from PIL import Image
+
+logger = logging.getLogger(__name__)
+logger.setLevel(logging.INFO)
+
+
+context.set_context(mode=context.GRAPH_MODE, device_target="Ascend")
+
+
+class LeNet5(nn.Cell):
+ """Lenet network structure."""
+
+ # define the operator required
+ def __init__(self, num_class=10, num_channel=1):
+ super(LeNet5, self).__init__()
+ self.conv1 = nn.Conv2d(num_channel, 6, 5, pad_mode='valid')
+ self.conv2 = nn.Conv2d(6, 16, 5, pad_mode='valid')
+ self.fc1 = nn.Dense(16 * 5 * 5, 120, weight_init=Normal(0.02))
+ self.fc2 = nn.Dense(120, 84, weight_init=Normal(0.02))
+ self.fc3 = nn.Dense(84, num_class, weight_init=Normal(0.02))
+ self.relu = nn.ReLU()
+ self.max_pool2d = nn.MaxPool2d(kernel_size=2, stride=2)
+ self.flatten = nn.Flatten()
+
+ # use the preceding operators to construct networks
+ def construct(self, x):
+ x = self.max_pool2d(self.relu(self.conv1(x)))
+ x = self.max_pool2d(self.relu(self.conv2(x)))
+ x = self.flatten(x)
+ x = self.relu(self.fc1(x))
+ x = self.relu(self.fc2(x))
+ x = self.fc3(x)
+ return x
+
+
+class MnistService(SingleNodeService):
+ def __init__(self, model_name, model_path):
+ self.model_name = model_name
+ self.model_path = model_path
+ logger.info("self.model_name:%s self.model_path: %s", self.model_name,
+ self.model_path)
+ self.network = None
+ # Load the model in non-blocking mode to prevent blocking timeout.
+ thread = threading.Thread(target=self.load_model)
+ thread.start()
+
+ def load_model(self):
+ logger.info("load network ... \n")
+ self.network = LeNet5()
+ ckpt_file = self.model_path + "/checkpoint_lenet_1-1_1875.ckpt"
+ logger.info("ckpt_file: %s", ckpt_file)
+ param_dict = load_checkpoint(ckpt_file)
+ load_param_into_net(self.network, param_dict)
+ # Inference warm-up. Otherwise, the initial inference will take a long time.
+ self.network_warmup()
+ logger.info("load network successfully ! \n")
+
+ def network_warmup(self):
+ # Inference warm-up. Otherwise, the initial inference will take a long time.
+ logger.info("warmup network ... \n")
+ images = np.array(np.random.randn(1, 1, 32, 32), dtype=np.float32)
+ inputs = Tensor(images, mindspore.float32)
+ inference_result = self.network(inputs)
+ logger.info("warmup network successfully ! \n")
+
+ def _preprocess(self, input_data):
+ preprocessed_result = {}
+ images = []
+ for k, v in input_data.items():
+ for file_name, file_content in v.items():
+ image1 = Image.open(file_content)
+ image1 = image1.resize((1, 32 * 32))
+ image1 = np.array(image1, dtype=np.float32)
+ images.append(image1)
+
+ images = np.array(images, dtype=np.float32)
+ logger.info(images.shape)
+ images.resize([len(input_data), 1, 32, 32])
+ logger.info("images shape: %s", images.shape)
+ inputs = Tensor(images, mindspore.float32)
+ preprocessed_result['images'] = inputs
+
+ return preprocessed_result
+
+ def _inference(self, preprocessed_result):
+ inference_result = self.network(preprocessed_result['images'])
+ return inference_result
+
+ def _postprocess(self, inference_result):
+ return str(inference_result)
+
+
+
+
+
diff --git a/docs/modelarts/umn/inference-modelarts-0078.html b/docs/modelarts/umn/inference-modelarts-0078.html
new file mode 100644
index 00000000..e1a28843
--- /dev/null
+++ b/docs/modelarts/umn/inference-modelarts-0078.html
@@ -0,0 +1,23 @@
+
+
+
+
Examples of Custom Scripts
+
+
+
+
+
diff --git a/docs/modelarts/umn/modelarts_23_0173.html b/docs/modelarts/umn/inference-modelarts-0079.html
similarity index 95%
rename from docs/modelarts/umn/modelarts_23_0173.html
rename to docs/modelarts/umn/inference-modelarts-0079.html
index dc0de58c..5f627de3 100644
--- a/docs/modelarts/umn/modelarts_23_0173.html
+++ b/docs/modelarts/umn/inference-modelarts-0079.html
@@ -1,8 +1,8 @@
-
+
TensorFlow
-
TensorFlow has two types of APIs: Keras and tf. Keras and tf use different code for training and saving models, but the same code for inference.
-
Training a Model (Keras API)
1
+There are two types of TensorFlow APIs, Keras and tf. They use different code for training and saving models, but the same code for inference.
+ Training a Model (Keras API) 1
2
3
4
@@ -75,7 +75,7 @@
Saving a Model (Keras API) 1
+Saving a Model (Keras API) 1
2
3
4
@@ -140,7 +140,7 @@
Training a Model (tf API) 1
+Training a Model (tf API) 1
2
3
4
@@ -593,7 +593,7 @@
Saving a Model (tf API) 1
+Saving a Model (tf API) 1
2
3
4
@@ -656,8 +656,8 @@
Inference Code (Keras and tf APIs)Inference code must be inherited from the BaseService class. For details about the import statements of different types of parent model classes, see Table 1.
- 1
+Inference Code (Keras and tf APIs)In the model inference code file customize_service.py, add a child model class which inherits properties from its parent model class. For details about the import statements of different types of parent model classes, see Table 1.
+ 1
2
3
4
@@ -699,7 +699,7 @@
from model_service.tfserving_model_service import TfServingBaseService
-class mnist_service(TfServingBaseService):
+class MnistService(TfServingBaseService):
# Match the model input with the user's HTTPS API input during preprocessing.
# The model input corresponding to the preceding training part is {"images":<array>}.
@@ -739,7 +739,7 @@
PyTorch
-Training a Modelfrom __future__ import print_function
+Training a Modelfrom __future__ import print_function
import argparse
import torch
import torch.nn as nn
@@ -82,11 +82,11 @@ for epoch in range(1, 2 + 1):
train(model, device, train_loader, optimizer, epoch)
test(model, device, test_loader)
- Saving a Model# The model must be saved using state_dict and can be deployed remotely.
+Saving a Model# The model must be saved using state_dict and can be deployed remotely.
torch.save(model.state_dict(), "pytorch_mnist/mnist_mlp.pt")
Inference CodeInference code must be inherited from the BaseService class. For details about the import statements of different types of parent model classes, see Table 1.
- from PIL import Image
+Inference CodeIn the model inference code file customize_service.py, add a child model class which inherits properties from its parent model class. For details about the import statements of different types of parent model classes, see Table 1.
+ from PIL import Image
import log
from model_service.pytorch_model_service import PTServingBaseService
import torch.nn.functional as F
@@ -122,7 +122,7 @@ class PTVisionService(PTServingBaseService):
# Load tags.
self.label = [0,1,2,3,4,5,6,7,8,9]
# Labels can also be loaded by label file.
- # Store the label.json file in the model directory. The following information is read:
+ # Store the label.json file in the model directory. The following information is read:
dir_path = os.path.dirname(os.path.realpath(self.model_path))
with open(os.path.join(dir_path, 'label.json')) as f:
self.label = json.load(f)
@@ -199,7 +199,7 @@ def Mnist(model_path, **kwargs):
Caffe
-Training and Saving a Modellenet_train_test.prototxt file
- 1
+Training and Saving a Modellenet_train_test.prototxt file
+ 1
2
3
4
@@ -340,8 +340,8 @@
|
- lenet_solver.prototxt file
- 1
+lenet_solver.prototxt file
+ 1
2
3
4
@@ -393,10 +393,10 @@
|
Train the model.
-./build/tools/caffe train --solver=examples/mnist/lenet_solver.prototxt
-The caffemodel file is generated after model training. Rewrite the lenet_train_test.prototxt file to the lenet_deploy.prototxt file used for deployment by modifying input and output layers.
- 1
+Train the model.
+./build/tools/caffe train --solver=examples/mnist/lenet_solver.prototxt
+The caffemodel file is generated after model training. Rewrite the lenet_train_test.prototxt file to the lenet_deploy.prototxt file used for deployment by modifying input and output layers.
+ 1
2
3
4
@@ -657,8 +657,8 @@
Inference CodeInference code must be inherited from the BaseService class. For details about the import statements of different types of parent model classes, see Table 1.
- 1
+Inference CodeIn the model inference code file customize_service.py, add a child model class which inherits properties from its parent model class. For details about the import statements of different types of parent model classes, see Table 1.
+ 1
2
3
4
@@ -762,7 +762,7 @@
Upgrading a Service
+For a deployed service, you can modify its basic information to match service changes and change the AI application version to upgrade it.
+ You can modify the basic information about a service in either of the following ways:
+ Method 1: Modify Service Information on the Service Management Page
+ Method 2: Modify Service Information on the Service Details Page
+ PrerequisitesThe service has been deployed. The service in the Deploying state cannot be upgraded by modifying the service information.
+
+ Constraints- Improper upgrade operations will interrupt service running during the upgrade. Therefore, exercise caution when performing this operation.
- ModelArts supports hitless rolling upgrade of real-time services in some scenarios. Before upgrade, prepare for it and confirm the prerequisites.
+
Table 1 Scenarios for hitless rolling upgradeMeta Model Source for Creating an AI Application
+ |
+Using a Public Resource Pool
+ |
+Using a Dedicated Resource Pool
+ |
+
|---|
+
+Training job
+ |
+Not supported
+ |
+Not supported
+ |
+
+Container image
+ |
+Not supported
+ |
+Supported. The custom image for creating an AI application must meet Custom Image Specifications for Creating AI Applications.
+ |
+
+OBS
+ |
+Not supported
+ |
+Not supported
+ |
+
+
+
+
+
+ Method 1: Modify Service Information on the Service Management Page- Log in to the ModelArts management console and choose Service Deployment from the left navigation pane. Go to the service management page of the target service.
- In the service list, click Modify in the Operation column of the target service, modify basic service information, and submit the modification task as prompted.
When some parameters are modified, the system automatically restarts the service for the modification to take effect. When you submit a service modification task, if a restart is required, a dialog box will be displayed.
+
+
+
+ Method 2: Modify Service Information on the Service Details Page- Log in to the ModelArts management console and choose Service Deployment from the left navigation pane. Go to the service management page of the target service.
- Click the name of the target service. The service details page is displayed.
- Click Modify in the upper right corner of the page, modify the service details, and submit the modification task as prompted.
When some parameters are modified, the system automatically restarts the service for the modification to take effect. When you submit a service modification task, if a restart is required, a dialog box will be displayed.
+
+
+
+ Starting, Stopping, Deleting, or Restarting a Service
+Starting a ServiceYou can start services in the Successful, Abnormal, or Stopped status. Services in the Deploying state cannot be started. You can start a service in the following ways:
+
+ - Log in to the ModelArts management console and choose Service Deployment from the left navigation pane. Go to the service management page of the target service. Click Start in the Operation column to start a service. (For a real-time service, choose More > Start in the Operation column.)
- Log in to the ModelArts management console and choose Service Deployment from the left navigation pane. Go to the service management page of the target service. Click the name of the target service. The service details page is displayed. Click Start in the upper right corner of the page to start the service.
+ Stopping a Service Stop a service in either of the following ways:
+ - Log in to the ModelArts management console and choose Service Deployment from the left navigation pane. Go to the service management page of the target service. Click Stop in the Operation column to stop a service. (For a real-time service, choose More > Stop in the Operation column.)
- Log in to the ModelArts management console and choose Service Deployment from the left navigation pane. Go to the service management page of the target service. Click the name of the target service. The service details page is displayed. Click Stop in the upper right corner of the page to stop the service.
+
+ Deleting a Service
+If a service is no longer in use, you can delete it to release resources.
+ Log in to the ModelArts management console and choose Service Deployment from the left navigation pane. Go to the service management page of the target service. - For a real-time service, choose More > Delete in the Operation column to delete the service
- For a batch service, click Delete in the Operation column to delete the service.
+ - A deleted service cannot be recovered.
- A service cannot be deleted without agency authorization.
+
+
+ What Is ModelArts?
-ModelArts is a one-stop AI development platform geared toward developers and data scientists of all skill levels. It enables you to rapidly build, train, and deploy models anywhere, and manage full-lifecycle AI workflows. ModelArts accelerates AI development and fosters AI innovation with key capabilities, including data preprocessing and auto labeling, distributed training, automated model building, and one-click workflow execution.
- ModelArts covers all stages of AI development, including data processing and model training and deployment. The underlying technologies of ModelArts support various heterogeneous computing resources, allowing developers to flexibly select and use resources. In addition, ModelArts supports popular open-source AI development frameworks such as TensorFlow and PyTorch. ModelArts also allows you to use customized algorithm frameworks tailored to your needs.
- ModelArts aims to simplify AI development.
- Product ArchitectureModelArts supports the entire development process, including data processing, and model training, management, and deployment.
- ModelArts supports various AI application scenarios, such as image classification and object detection.
+ ModelArts is a one-stop AI development platform geared toward developers and data scientists of all skill levels. It enables you to rapidly build, train, and deploy models anywhere, and manage full-lifecycle AI workflows. ModelArts accelerates AI development and fosters AI innovation with key capabilities, including data preprocessing and auto labeling, distributed training, automated model building, and one-click workflow execution.
+ ModelArts covers all stages of AI development, including data processing, algorithm development, and model training and deployment. The underlying technologies of ModelArts support various heterogeneous computing resources, allowing developers to flexibly select and use resources. In addition, ModelArts supports popular open-source AI development frameworks such as TensorFlow, PyTorch, and MindSpore. ModelArts also allows you to use customized algorithm frameworks tailored to your needs.
+ ModelArts aims to simplify AI development.
+ Product ArchitectureModelArts supports the entire development process, including data processing, and model training, management, and deployment. It also provides AI Gallery for sharing models.
+ ModelArts supports various AI application scenarios, such as image classification, object detection, video analysis, speech recognition, product recommendation, and exception detection.
- Figure 1 ModelArts architecture
-
+
+
\ No newline at end of file
diff --git a/docs/modelarts/umn/modelarts_01_0003.html b/docs/modelarts/umn/modelarts_01_0003.html
index b6004f9a..274d0d39 100644
--- a/docs/modelarts/umn/modelarts_01_0003.html
+++ b/docs/modelarts/umn/modelarts_01_0003.html
@@ -1,19 +1,26 @@
-
+
Functions
- AI engineers face challenges in the installation and configuration of various AI tools, data preparation, and model training. To address these challenges, the one-stop AI development platform ModelArts is provided. The platform integrates data preparation, algorithm development, model training, and model deployment into the production environment, allowing AI engineers to perform one-stop AI development.
- Figure 1 Function overview
- ModelArts has the following features:
- - Data governance
Manages data preparation, such as data filtering and labeling, and dataset versions.
- - Rapid and simplified model training
Enables high-performance distributed training and simplifies coding with the self-developed MoXing deep learning framework.
- - Multi-scenario deployment
Deploys models in various production environments, and supports real-time and batch inference.
- - Auto learning
Enables model building without coding and supports image classification, object detection, and predictive analytics.
+AI engineers face challenges in the installation and configuration of various AI tools, data preparation, and model training. To address these challenges, the one-stop AI development platform ModelArts is provided. The platform integrates data preparation, algorithm development, model training, and model deployment into the production environment, allowing AI engineers to perform one-stop AI development.
+ Figure 1 Function overview
+ ModelArts has the following features:
+ - Data governance
Manages data preparation, such as data filtering and labeling, and dataset versions.
+ - Rapid and simplified model training
Enables high-performance distributed training and simplifies coding with the self-developed MoXing deep learning framework.
+ - Multi-scenario deployment
Deploys models in various production environments, and supports real-time and batch inference.
+ - Auto learning
Enables model building without coding and supports image classification, object detection, and predictive analytics.
Common Concepts of ModelArts
-ExeMLExeML is the process of automating model design, parameter tuning, and model training, model compression, and model deployment with the labeled data. The process is code-free and does not require developers to have experience in model development. A model can be built in three steps: labeling data, training a model, and deploying the model.
+ ExeMLExeML is the process of automating model design, parameter tuning, and model training, model compression, and model deployment with the labeled data. The process is code-free and does not require developers to have experience in model development. A model can be built in three steps: labeling data, training a model, and deploying the model.
- InferenceInference is the process of deriving a new judgment from a known judgment according to a certain strategy. In AI, machines simulate human intelligence, and complete inference based on neural networks.
+ InferenceInference is the process of deriving a new judgment from a known judgment according to a certain strategy. In AI, machines simulate human intelligence, and complete inference based on neural networks.
- Real-Time InferenceReal-time inference specifies a web service that provides an inference result for each inference request.
+ Real-Time InferenceReal-time inference specifies a web service that provides an inference result for each inference request.
- Batch InferenceBatch inference specifies a batch job that processes batch data for inference.
+ Batch InferenceBatch inference specifies a batch job that processes batch data for inference.
- Ascend ChipThe Ascend chips are a series of AI chips with high computing performance and low power consumption.
+ Ascend ChipThe Ascend chips are a series of AI chips with high computing performance and low power consumption.
- Resource PoolModelArts provides large-scale computing clusters for model development, training, and deployment. There are two types of resource pools: public resource pool and dedicated resource pool. The public resource pool is provided by default. Dedicated resource pools are created separately and used exclusively.
+ Resource PoolModelArts provides large-scale computing clusters for model development, training, and deployment. There are two types of resource pools: public resource pool and dedicated resource pool. The public resource pool is provided by default. Dedicated resource pools are created separately and used exclusively.
diff --git a/docs/modelarts/umn/modelarts_01_0006.html b/docs/modelarts/umn/modelarts_01_0006.html
index 1b146496..5f7401c8 100644
--- a/docs/modelarts/umn/modelarts_01_0006.html
+++ b/docs/modelarts/umn/modelarts_01_0006.html
@@ -1,20 +1,24 @@
-
+
Related Services
- OBSModelArts uses Object Storage Service (OBS) to securely and reliably store data and models at low costs. For more details, see Object Storage Service Console Operation Guide.
+ IAMModelArts uses Identity and Access Management (IAM) for authentication and authorization. For more information about IAM, see Identity and Access Management User Guide.
- CCEModelArts uses Cloud Container Engine (CCE) to deploy models as real-time services. CCE enables high concurrency and provides elastic scaling. For more information about CCE, see Cloud Container Engine User Guide.
+ OBSModelArts uses Object Storage Service (OBS) to securely and reliably store data and models at low costs. For more details, see Object Storage Service Console Operation Guide.
- SWRTo use an AI framework that is not supported by ModelArts, use SoftWare Repository for Container (SWR) to customize an image and import the image to ModelArts for training or inference. For details about SWR, see .
+ EVSModelArts uses Elastic Volume Service (EVS) to store created notebook instances.
- Cloud EyeModelArts uses Cloud Eye to monitor online services and model loads in real time and send alarms and notifications automatically. For details about Cloud Eye, see Cloud Eye User Guide.
+ CCEModelArts uses Cloud Container Engine (CCE) to deploy models as real-time services. CCE enables high concurrency and provides elastic scaling. For more information about CCE, see Cloud Container Engine User Guide.
- CTSModelArts uses Cloud Trace Service (CTS) to record operations for later query, audit, and backtrack operations. For details about CTS, see Cloud Trace Service User Guide.
+ SWRTo use an AI framework that is not supported by ModelArts, use Software Repository for Container (SWR) to customize an image and import the image to ModelArts for training or inference. For details about SWR, see .
+
+ Cloud EyeModelArts uses Cloud Eye to monitor online services and model loads in real time and send alarms and notifications automatically. For details about Cloud Eye, see Cloud Eye User Guide.
+
+ CTSModelArts uses Cloud Trace Service (CTS) to record operations for later query, audit, and backtrack operations. For details about CTS, see Cloud Trace Service User Guide.
diff --git a/docs/modelarts/umn/modelarts_01_0007.html b/docs/modelarts/umn/modelarts_01_0007.html
index a9cddee1..23c7cd20 100644
--- a/docs/modelarts/umn/modelarts_01_0007.html
+++ b/docs/modelarts/umn/modelarts_01_0007.html
@@ -1,14 +1,14 @@
-
+
How Do I Access ModelArts?
- You can access ModelArts through the web-based management console or by using HTTPS-based application programming interfaces (APIs).
- - Using the Management Console
ModelArts features a simple and easy-to-use management console, and provides a host of functions including ExeML, data management, development environment, model training, model management, and service deployment. You can complete end-to-end AI development on the management console.
- - Using APIs
If you want to integrate ModelArts into a third-party system for secondary development, use APIs to access ModelArts. For details about the APIs and operations, see ModelArts API Reference.
+You can access ModelArts through the web-based management console or by using HTTPS-based application programming interfaces (APIs).
+ - Using the Management Console
ModelArts features a simple and easy-to-use management console, and provides a host of functions including ExeML, data management, development environment, model training, AI application management, AI Hub, and service deployment. You can complete end-to-end AI development on the management console.
+ - Using APIs
If you want to integrate ModelArts into a third-party system for secondary development, use APIs to access ModelArts. For details about the APIs and operations, see ModelArts API Reference.
Basic Knowledge
@@ -12,20 +12,18 @@
- Data Management
-- DevEnviron (Old Version)
-
-- Introduction to Development Tools
+ - Introduction to Development Tools
- Model Training
- Model Deployment
-- ExeML
+ - ExeML
diff --git a/docs/modelarts/umn/modelarts_01_0010.html b/docs/modelarts/umn/modelarts_01_0010.html
index 7961e6fd..e63e3f12 100644
--- a/docs/modelarts/umn/modelarts_01_0010.html
+++ b/docs/modelarts/umn/modelarts_01_0010.html
@@ -1,20 +1,20 @@
-
+
Introduction to the AI Development Lifecycle
- What Is AIArtificial intelligence (AI) is a technology capable of simulating human cognition through machines. The core capability of AI is to make a judgment or prediction based on a given input.
+ What Is AIArtificial intelligence (AI) is a technology capable of simulating human cognition through machines. The core capability of AI is to make a judgment or prediction based on a given input.
- What Is the Purpose of AI DevelopmentAI development aims to centrally process and extract information from volumes of data to summarize internal patterns of the study objects.
- Massive volumes of collected data are computed, analyzed, summarized, and organized by using appropriate statistics, machine learning, and deep learning methods to maximize data value.
+ What Is the Purpose of AI DevelopmentAI development aims to centrally process and extract information from volumes of data to summarize internal patterns of the study objects.
+ Massive volumes of collected data are computed, analyzed, summarized, and organized by using appropriate statistics, machine learning, and deep learning methods to maximize data value.
- Basic Process of AI DevelopmentThe basic process of AI development includes the following steps: determining an objective, preparing data, and training, evaluating, and deploying a model.
- Figure 1 AI development process
- - Determine an objective.
Before starting AI development, determine what to analyze. What problems do you want to solve? What is the business goal? Sort out the AI development framework and ideas based on the business understanding. For example, image classification and object detection. Different projects have different requirements for data and AI development methods.
- - Prepare data.
Data preparation refers to data collection and preprocessing.
-Data preparation is the basis of AI development. When you collect and integrate related data based on the determined objective, the most important thing is to ensure the authenticity and reliability of the obtained data. Typically, you cannot collect all the data at the same time. In the data labeling phase, you may find that some data sources are missing and then you may need to repeatedly adjust and optimize the data.
- - Train a model.
Modeling involves analyzing the prepared data to find the causality, internal relationships, and regular patterns, thereby providing references for commercial decision making. After model training, usually one or more machine learning or deep learning models are generated. These models can be applied to new data to obtain predictions and evaluation results.
- - Evaluate the model.
A model generated by training needs to be evaluated. Typically, you cannot obtain a satisfactory model after the first evaluation, and may need to repeatedly adjust algorithm parameters and data to further optimize the model.
-Some common metrics, such as the accuracy, recall, and area under the curve (AUC), help you effectively evaluate and obtain a satisfactory model.
- - Deploy the model.
Model development and training are based on existing data (which may be test data). After a satisfactory model is obtained, the model needs to be formally applied to actual data or newly generated data for prediction, evaluation, and visualization. The findings can then be reported to decision makers in an intuitive way, helping them develop the right business strategies.
+Basic Process of AI DevelopmentThe basic process of AI development includes the following steps: determining an objective, preparing data, and training, evaluating, and deploying a model.
+ Figure 1 AI development process
+ - Determine an objective.
Before starting AI development, determine what to analyze. What problems do you want to solve? What is the business goal? Sort out the AI development framework and ideas based on the business understanding. For example, image classification and object detection. Different projects have different requirements for data and AI development methods.
+ - Prepare data.
Data preparation refers to data collection and preprocessing.
+Data preparation is the basis of AI development. When you collect and integrate related data based on the determined objective, the most important thing is to ensure the authenticity and reliability of the obtained data. Typically, you cannot collect all the data at the same time. In the data labeling phase, you may find that some data sources are missing and then you may need to repeatedly adjust and optimize the data.
+ - Train a model.
Modeling involves analyzing the prepared data to find the causality, internal relationships, and regular patterns, thereby providing references for commercial decision making. After model training, usually one or more machine learning or deep learning models are generated. These models can be applied to new data to obtain predictions and evaluation results.
+ - Evaluate the model.
A model generated by training needs to be evaluated. Typically, you cannot obtain a satisfactory model after the first evaluation, and may need to repeatedly adjust algorithm parameters and data to further optimize the model.
+Some common metrics, such as the accuracy, recall, and area under the curve (AUC), help you effectively evaluate and obtain a satisfactory model.
+ - Deploy the model.
Model development and training are based on existing data (which may be test data). After a satisfactory model is obtained, the model needs to be formally applied to actual data or newly generated data for prediction, evaluation, and visualization. The findings can then be reported to decision makers in an intuitive way, helping them develop the right business strategies.
@@ -24,3 +24,10 @@
+
+
\ No newline at end of file
diff --git a/docs/modelarts/umn/modelarts_01_0011.html b/docs/modelarts/umn/modelarts_01_0011.html
index 7f464703..de96ea03 100644
--- a/docs/modelarts/umn/modelarts_01_0011.html
+++ b/docs/modelarts/umn/modelarts_01_0011.html
@@ -1,17 +1,17 @@
-
+
Basic Concepts of AI Development
- Machine learning is classified into supervised, unsupervised, and reinforcement learning.
- - Supervised learning uses labeled samples to adjust the parameters of classifiers to achieve the required performance. It can be considered as learning with a teacher. Common supervised learning includes regression and classification.
- Unsupervised learning is used to find hidden structures in unlabeled data. Clustering is a form of unsupervised learning.
- Reinforcement learning is an area of machine learning concerned with how software agents ought to take actions in an environment so as to maximize some notion of cumulative reward.
- RegressionRegression reflects the time feature of data attributes and generates a function that maps one data attribute to an actual variable prediction to find the dependency between the variable and attribute. Regression mainly analyzes data and predicts data and data relationship. Regression can be used for customer development, retention, customer churn prevention, production lifecycle analysis, sales trend prediction, and targeted promotion.
- 
+ Machine learning is classified into supervised, unsupervised, and reinforcement learning.
+ - Supervised learning uses labeled samples to adjust the parameters of classifiers to achieve the required performance. It can be considered as learning with a teacher. Common supervised learning includes regression and classification.
- Unsupervised learning is used to find hidden structures in unlabeled data. Clustering is a form of unsupervised learning.
- Reinforcement learning is an area of machine learning concerned with how software agents ought to take actions in an environment so as to maximize some notion of cumulative reward.
+ RegressionRegression reflects the time feature of data attributes and generates a function that maps one data attribute to an actual variable prediction to find the dependency between the variable and attribute. Regression mainly analyzes data and predicts data and data relationship. Regression can be used for customer development, retention, customer churn prevention, production lifecycle analysis, sales trend prediction, and targeted promotion.
+ 
- ClassificationClassification involves defining a set of categories based on the common features of objects and identifying which category an object belongs to. Classification can be used for customer classification, customer properties, feature analysis, customer satisfaction analysis, and customer purchase trend prediction.
- 
+ ClassificationClassification involves defining a set of categories based on the common features of objects and identifying which category an object belongs to. Classification can be used for customer classification, customer properties, feature analysis, customer satisfaction analysis, and customer purchase trend prediction.
+ 
- ClusteringClustering involves grouping a set of objects in such a way that objects in the same group are more similar to each other than to those in other groups. Clustering can be used for customer segmentation, customer characteristic analysis, customer purchase trend prediction, and market segmentation.
- 
- Clustering analyzes data objects and produces class labels. Objects are grouped based on the maximized and minimized similarities to form clusters. In this way, objects in the same cluster are more similar to each other than to those in other clusters.
+ ClusteringClustering involves grouping a set of objects in such a way that objects in the same group are more similar to each other than to those in other groups. Clustering can be used for customer segmentation, customer characteristic analysis, customer purchase trend prediction, and market segmentation.
+ 
+ Clustering analyzes data objects and produces class labels. Objects are grouped based on the maximized and minimized similarities to form clusters. In this way, objects in the same cluster are more similar to each other than to those in other clusters.
@@ -20,3 +20,10 @@
+
+
\ No newline at end of file
diff --git a/docs/modelarts/umn/modelarts_01_0012.html b/docs/modelarts/umn/modelarts_01_0012.html
index 5205e618..f4e411b4 100644
--- a/docs/modelarts/umn/modelarts_01_0012.html
+++ b/docs/modelarts/umn/modelarts_01_0012.html
@@ -1,11 +1,14 @@
-
+
Data Management
- During AI development, massive volumes of data need to be processed, and data preparation and labeling usually take more than half of the development time. ModelArts data management provides an efficient data management and labeling framework. It supports different data types such as image, text, audio, and video, and covers a range of labeling scenarios such as image classification, object detection, speech paragraph labeling, and text classification. It is used to AI projects such as computer vision, natural language processing, and audio and video analysis. In addition, it provides functions such as data filtering, data analysis, data processing, team labeling, and version management. This framework enables AI developers to fulfill the entire data labeling process. See the following figure.
- Figure 1 Data labeling process
- You can select the appropriate labeling tool in data management to label data in the specified scenario. Models trained by built-in algorithms or custom algorithms can be selected for intelligent labeling. Only a small amount of manual labeling and correction are required to obtain accurate labeling results. You can create a team to perform collaborative labeling, improving labeling efficiency. ModelArts allows project-based management for labeling by individual developers, small-scale labeling by small teams, and large-scale labeling by professional teams.
- For large-scale team labeling, ModelArts provides team management, personnel management, and data management to implement the entire process, from project creation, allocation, management, labeling, to acceptance. For small-scale labeling by individuals and small teams, ModelArts provides an easy-to-use labeling tool to minimize project management costs.
- In addition, the labeling platform ensures data security. User data is used only within the authorized scope. The labeling object allocation policy ensures user data privacy and implements data anonymization.
+ During AI development, massive volumes of data need to be processed, and data preparing and labeling usually take more than half of the time required for the entire development process. ModelArts data management provides an efficient data management and labeling framework. It supports image, text, audio, and video data types in a range of labeling scenarios such as image classification, object detection, speech paragraph labeling, and text classification so that data management can be used in various AI projects such as computer vision, natural language processing, and audio and video analysis projects. In addition, ModelArts data management provides functions such as data filtering, data analysis, data processing, team labeling, and version management, enabling you to manage the full data labeling process. Figure 1 shows the data labeling process.
+ Figure 1 Data labeling process
+ ModelArts data management analyzes and processes data using such functions as clustering analysis, data feature analysis, data cleansing, data verification, data augmentation, and data selection, helping you obtain high-value data that meets development or project requirements.
+ With data management, ModelArts allows you to label data online for image classification, object detection, speech paragraphs, text triplet, and videos. You can also use intelligent labeling to automatically label data through built-in or customized algorithms, improving the labeling efficiency.
+ To support large-scale collaborative labeling, data management provides team labeling with team management, personnel management, and data management for full-process project management, from project creation, data allocation, progress control, labeling, review, to acceptance. This improves labeling efficiency and minimizes project management costs.
+ ModelArts data management ensures the security and privacy of user data and allows data to be used only within the authorized scope.
+ In the new version of data management, datasets and data labeling are decoupled to facilitate your operations.
+
+
+
\ No newline at end of file
diff --git a/docs/modelarts/umn/modelarts_01_0013-.html b/docs/modelarts/umn/modelarts_01_0013-.html
deleted file mode 100644
index 8210bc39..00000000
--- a/docs/modelarts/umn/modelarts_01_0013-.html
+++ /dev/null
@@ -1,15 +0,0 @@
-
-
- DevEnviron (Old Version)
- It is challenging to set up a development environment, select an AI algorithm framework and algorithm, debug code, install software, and accelerate hardware. To help users overcome these challenges, ModelArts simplifies the entire development process and lowers the development threshold. Figure 1 shows the algorithm development process.
- Figure 1 Algorithm development
- - Support for popular AI algorithm frameworks
In the machine learning and deep learning fields, popular open-source training and inference frameworks include TensorFlow, PyTorch, MXNet, and MindSpore. ModelArts supports popular AI computing frameworks and provides a user-friendly development and debugging environment. It supports traditional machine learning algorithms, such as logistic regression, decision tree, and clustering, as well as multiple types of deep learning algorithms, such as the convolutional neural network (CNN), recurrent neural network (RNN), and long short-term memory (LSTM).
- - Simplified algorithm development for distributed training
Deep learning generally requires large-scale GPU clusters for distributed acceleration. For existing open-source frameworks, algorithm developers need to write a large amount of code for distributed training on different hardware, and the acceleration code varies depending on the framework. To resolve these issues, a distributed lightweight framework or SDK is required. The framework or SDK is built on deep learning engines such as TensorFlow, PyTorch, MXNet, and MindSpore to improve the distributed performance and usability of these engines. ModelArts MoXing perfectly suits the needs. The easy-to-use MoXing API/SDK enables you to develop deep learning at low costs.
-
-
-
-
diff --git a/docs/modelarts/umn/modelarts_01_0013.html b/docs/modelarts/umn/modelarts_01_0013.html
deleted file mode 100644
index 3e8ba661..00000000
--- a/docs/modelarts/umn/modelarts_01_0013.html
+++ /dev/null
@@ -1,34 +0,0 @@
-
-
- Introduction to Development Tools
- This document describes the DevEnviron notebook functions of the new version.
-
- Software development is a process of reducing developer costs and improving development experience. In AI development, ModelArts is dedicated to improving AI development experience and simplifying the development process. ModelArts DevEnviron uses cloud native resources and integrates the development tool chain to provide better in-cloud AI development experience for AI development, exploration, and teaching.
- ModelArts notebook for seamless in-cloud and on-premises collaboration
- - In-cloud JupyterLab, local IDE, and ModelArts plug-ins for remote development and debugging, tailored to your needs
- In-cloud development environment with AI compute resources, cloud storage, and built-in AI engines
- Custom runtime environment saved as an image for training and inference
-
- Feature 1: Remote development, allowing remote access to notebook from a local IDEThe notebook of the new version provides remote development. After enabling remote SSH, you can remotely access the ModelArts notebook development environment to debug and run code from a local IDE.
- Due to limited local resources, developers using a local IDE run and debug code typically on a CPU or GPU server shared between team members. Building and maintaining the CPU or GPU server are costly.
- ModelArts notebook instances are out of the box with various built-in engines and flavors for you to select. You can use a dedicated container environment. Only after simple configurations, you can remotely access the environment to run and debug code from your local IDE.
- ModelArts notebook can be regarded as an extension of a local development environment. The operations such as data reading, training, and file saving are the same as those performed in a local environment.
- ModelArts notebook allows you to use in-cloud resources while with local coding habits unchanged.
- A local IDE supports Visual Studio (VS) Code, PyCharm, and SSH. In addition, the PyCharm Toolkit plug-ins allow you to easily use cloud resources.
-
- Feature 2: Preset images that are out-of-the-box with optimized configurations and supporting mainstream AI enginesThe AI engines and versions preset in each image are fixed. When creating a notebook instance, specify an AI engine and version, including the chip type.
- ModelArts DevEnviron provides a group of preset images, including PyTorch, TensorFlow, and MindSpore images. You can use a preset image to start your notebook instance. After the development in the instance, submit a training job without any adaptation.
- The image versions preset in ModelArts are determined based on user feedback and version stability. If your development can be carried out using the versions preset in ModelArts, for example, MindSpore 1.5, use preset images. These images have been fully verified and have many commonly-used installation packages built in. They are out-of-the-box, relieving you from configuring the environment.
- The images preset in ModelArts DevEnviron include:
- - Common preset packages: common AI engines such as PyTorch and MindSpore based on standard Conda, common data analysis software packages such as Pandas and Numpy, and common tool software such as CUDA and CUDNN, meeting common AI development requirements.
- Preset Conda environments: A Conda environment and basic Conda Python (excluding any AI engine) are created for each preset image. Select a Conda environment based on whether the AI engine is used for debugging.
- Notebook: a web application that enables you to code on the GUI and combine the code, mathematical equations, and visualized content into a document.
- JupyterLab plug-ins: enable flavor changing and instance stopping to improving user experience.
- Remote SSH: allows you to remotely debug a notebook instance from a local PC.
- - To simplify operations, ModelArts notebook of the new version does not support switchover between AI engines in a notebook instance.
- AI engines vary based on regions. For details about the AI engines available in a region, see the AI engines displayed on the management console.
-
-
- Feature 3: JupyterLab, an online interactive development and debugging toolModelArts integrates open-source JupyterLab for online interactive development and debugging. You can use the notebook on the ModelArts management console to compile and debug code and train models based on the code, without concerning environment installation or configuration.
- JupyterLab is an interactive development environment. It is the next-generation product of Jupyter Notebook. JupyterLab enables you to compile notebooks, operate terminals, edit Markdown text, enable interaction, and view CSV files and images.
-
-
-
-
diff --git a/docs/modelarts/umn/modelarts_01_0014.html b/docs/modelarts/umn/modelarts_01_0014.html
index 36cbec01..525579ea 100644
--- a/docs/modelarts/umn/modelarts_01_0014.html
+++ b/docs/modelarts/umn/modelarts_01_0014.html
@@ -1,18 +1,18 @@
-
+
Model Training
- In addition to data and algorithms, developers spend a lot of time configuring model training parameters. Model training parameters determine the model's precision and convergence time. Parameter selection is heavily dependent on developers' experience. Improper parameter selection will affect the model's precision or significantly increase the time required for model training.
- To simplify AI development and improve development efficiency and training performance, ModelArts offers visualized job management, resource management, and version management and automatically performs hyperparameter optimization based on machine learning and reinforcement learning. It provides automatic hyperparameter tuning policies such as learning rate and batch size, and integrates common models.
- Currently, when most developers build models, the models usually have dozens of layers or even hundreds of layers and MB-level or GB-level parameters to meet precision requirements. As a result, the specifications of computing resources are extremely high, especially the computing power of hardware resources, memory, and ROM. The resource specifications on the device side are strictly limited. For example, the computing power on the device side is 1 TFLOPS, the memory size is about 2 GB, and the ROM space is about 2 GB, so the model size on the device side must be limited to 100 KB and the inference delay must be limited to 100 milliseconds.
- Therefore, compression technologies with lossless or near-lossless model precision, such as pruning, quantization, and knowledge distillation, are used to implement automatic model compression and optimization, and automatic iteration of model compression and retraining to control the loss of model precision. The low-bit quantization technology, which eliminates the need for retraining, converts the model from a high-precision floating point to a fixed-point operation. Multiple compression and optimization technologies are used to meet the lightweight requirements of device and edge hardware resources. The model compression technology reduces the precision by less than 1% in specific scenarios.
- When the training data volume is large, the training of the deep learning model is time-consuming. In computer vision technology, ImageNet-1k (a classification dataset containing 1,000 image classes, referred to as ImageNet) is a commonly used dataset. If you use a P100 GPU to train a ResNet-50 model on the dataset, it will take nearly one week. This hinders rapid development of deep learning applications. Therefore, the acceleration of deep learning training has always been an important concern to the academia and the industry.
- Distributed training acceleration needs to be considered in terms of software and hardware. A single optimization method cannot meet expectations. Therefore, optimization of distributed acceleration is a system project. The distributed training architecture needs to be considered in terms of hardware and chip design. To minimize compute and communication delays, many factors need to be considered, including overall compute specifications, network bandwidth, high-speed cache, power consumption, and heat dissipation of the system, and the relationship between compute and communication throughput.
- The software design needs to combine high-performance hardware features to fully use the high-speed hardware network and implement high-bandwidth distributed communication and efficient local data caching. By using training optimization algorithms, such as hybrid parallel, gradient compression, and convolution acceleration, the software and hardware of the distributed training system can be efficiently coordinated and optimized from end to end, and training acceleration can be implemented in a distributed environment of multiple hosts and cards. ModelArts delivers an industry-leading speedup of over 0.8 for ResNet50 on the ImageNet dataset in the distributed environment with thousands of hosts and cards.
- To measure the acceleration performance of distributed deep learning, the following two key indicators are used:
- - Throughput, that is, the amount of data processed in a unit time
- Convergence time, that is, the time required to achieve certain precision
- The throughput depends on server hardware (for example, more AI acceleration chips with higher FLOPS processing capabilities and higher communication bandwidth achieve higher throughput), data reading and caching, data preprocessing, model computing (for example, convolution algorithm selection), and communication topology optimization. Except low-bit computing and gradient (or parameter) compression, most technologies improve throughput without affecting model precision. To achieve the shortest convergence time, you need to optimize the throughput and adjust the parameters. If the parameters are not adjusted properly, the throughput cannot be optimized. If the batch size is set to a small value, the parallel performance of model training will be relatively poor. As a result, the throughput cannot be improved even if the number of compute nodes are increased.
- Users are most concerned about convergence time. The MoXing framework implements full-stack optimization and significantly reduces the training convergence time. For data read and preprocessing, MoXing uses multi-level concurrent input pipelines to prevent data I/Os from becoming a bottleneck. In terms of model computing, MoXing provides hybrid precision calculation, which combines semi-precision and single-precision for the upper layer models and reduces the loss caused by precision calculation through adaptive scaling. Dynamic hyperparameter policies (such as momentum and batch size) are used to minimize the number of epochs required for model convergence.
- ModelArts High-Performance Distributed Training Optimization- Automatic hybrid precision to fully utilize hardware computing capabilities
- Dynamic hyperparameter adjustment technologies (dynamic batch size, image size, and momentum)
- Automatic model gradient merging and splitting
- Communication operator scheduling optimization based on BP bubble adaptive computing
- Distributed high-performance communication libraries (NStack and HCCL)
- Distributed data-model hybrid parallel
- Training data compression and multi-level caching
+ In addition to data and algorithms, developers spend a lot of time configuring model training parameters. Model training parameters determine the model's precision and convergence time. Parameter selection is heavily dependent on developers' experience. Improper parameter selection will affect the model's precision or significantly increase the time required for model training.
+ To simplify AI development and improve development efficiency and training performance, ModelArts offers visualized job management, resource management, and version management and automatically performs hyperparameter optimization based on machine learning and reinforcement learning. It provides automatic hyperparameter tuning policies such as learning rate and batch size, and integrates common models.
+ Currently, when most developers build models, the models usually have dozens of layers or even hundreds of layers and MB-level or GB-level parameters to meet precision requirements. As a result, the specifications of computing resources are extremely high, especially the computing power of hardware resources, memory, and ROM. The resource specifications on the device side are strictly limited. For example, the computing power on the device side is 1 TFLOPS, the memory size is about 2 GB, and the ROM space is about 2 GB, so the model size on the device side must be limited to 100 KB and the inference delay must be limited to 100 milliseconds.
+ Therefore, compression technologies with lossless or near-lossless model precision, such as pruning, quantization, and knowledge distillation, are used to implement automatic model compression and optimization, and automatic iteration of model compression and retraining to control the loss of model precision. The low-bit quantization technology, which eliminates the need for retraining, converts the model from a high-precision floating point to a fixed-point operation. Multiple compression and optimization technologies are used to meet the lightweight requirements of device and edge hardware resources. The model compression technology reduces the precision by less than 1% in specific scenarios.
+ When the training data volume is large, the training of the deep learning model is time-consuming. In computer vision technology, ImageNet-1k (a classification dataset containing 1,000 image classes, referred to as ImageNet) is a commonly used dataset. If you use a P100 GPU to train a ResNet-50 model on the dataset, it will take nearly one week. This hinders rapid development of deep learning applications. Therefore, the acceleration of deep learning training has always been an important concern to the academia and the industry.
+ Distributed training acceleration needs to be considered in terms of software and hardware. A single optimization method cannot meet expectations. Therefore, optimization of distributed acceleration is a system project. The distributed training architecture needs to be considered in terms of hardware and chip design. To minimize compute and communication delays, many factors need to be considered, including overall compute specifications, network bandwidth, high-speed cache, power consumption, and heat dissipation of the system, and the relationship between compute and communication throughput.
+ The software design needs to combine high-performance hardware features to fully use the high-speed hardware network and implement high-bandwidth distributed communication and efficient local data caching. By using training optimization algorithms, such as hybrid parallel, gradient compression, and convolution acceleration, the software and hardware of the distributed training system can be efficiently coordinated and optimized from end to end, and training acceleration can be implemented in a distributed environment of multiple hosts and cards. ModelArts delivers an industry-leading speedup of over 0.8 for ResNet50 on the ImageNet dataset in the distributed environment with thousands of hosts and cards.
+ To measure the acceleration performance of distributed deep learning, the following two key indicators are used:
+ - Throughput, that is, the amount of data processed in a unit time
- Convergence time, that is, the time required to achieve certain precision
+ The throughput depends on server hardware (for example, more AI acceleration chips with higher FLOPS processing capabilities and higher communication bandwidth achieve higher throughput), data reading and caching, data preprocessing, model computing (for example, convolution algorithm selection), and communication topology optimization. Except low-bit computing and gradient (or parameter) compression, most technologies improve throughput without affecting model precision. To achieve the shortest convergence time, you need to optimize the throughput and adjust the parameters. If the parameters are not adjusted properly, the throughput cannot be optimized. If the batch size is set to a small value, the parallel performance of model training will be relatively poor. As a result, the throughput cannot be improved even if the number of compute nodes are increased.
+ Users are most concerned about convergence time. The MoXing framework implements full-stack optimization and significantly reduces the training convergence time. For data read and preprocessing, MoXing uses multi-level concurrent input pipelines to prevent data I/Os from becoming a bottleneck. In terms of model computing, MoXing provides hybrid precision calculation, which combines semi-precision and single-precision for the upper layer models and reduces the loss caused by precision calculation through adaptive scaling. Dynamic hyperparameter policies (such as momentum and batch size) are used to minimize the number of epochs required for model convergence.
+ ModelArts High-Performance Distributed Training Optimization- Automatic hybrid precision to fully utilize hardware computing capabilities
- Dynamic hyperparameter adjustment technologies (dynamic batch size, image size, and momentum)
- Automatic model gradient merging and splitting
- Communication operator scheduling optimization based on BP bubble adaptive computing
- Distributed high-performance communication libraries (NStack and HCCL)
- Distributed data-model hybrid parallel
- Training data compression and multi-level caching
diff --git a/docs/modelarts/umn/modelarts_01_0015.html b/docs/modelarts/umn/modelarts_01_0015.html
index 7a233206..13be0455 100644
--- a/docs/modelarts/umn/modelarts_01_0015.html
+++ b/docs/modelarts/umn/modelarts_01_0015.html
@@ -1,10 +1,10 @@
-
+
Model Deployment
-
- Generally, AI model deployment and large-scale implementation are complex.
- Figure 1 Process of deploying a model
- - The real-time inference service features high concurrency, low latency, and elastic scaling, and supports multi-model gray release and A/B testing.
- ModelArts is optimized based on the high-performance AI inference chip Ascend 310. It can process PBs of inference data within a single day, publish over 1 million inference APIs on the cloud, and control inference network latency to milliseconds.
+ ModelArts is capable of managing models and services. This allows mainstream framework images and models from multiple vendors to be managed in a unified manner.
+ Generally, AI model deployment and large-scale implementation are complex.
+ Figure 1 Process of deploying a model
+ - The real-time inference service features high concurrency, low latency, and elastic scaling, and supports multi-model gray release and A/B testing.
- ModelArts is optimized based on the high-performance AI inference chip Ascend 310. It can process PBs of inference data within a single day, publish over 1 million inference APIs on the cloud, and control inference network latency to milliseconds.
+
+
\ No newline at end of file
diff --git a/docs/modelarts/umn/modelarts_01_0016.html b/docs/modelarts/umn/modelarts_01_0016.html
new file mode 100644
index 00000000..fb8b3e2b
--- /dev/null
+++ b/docs/modelarts/umn/modelarts_01_0016.html
@@ -0,0 +1,22 @@
+
+
+ ExeML
+ To implement AI in various industries, AI model development must be simplified. Currently, only a few algorithm engineers and researchers are capable of AI development and optimization. They find it challenging to develop related prototypes into products and projects. Most service developers, however, face difficulties in developing AI algorithms and optimizing parameters. As a result, most enterprises lack comprehensive AI development capabilities.
+ ModelArts uses machine learning to help service developers who are not experienced in algorithm development to develop algorithms. It automatically generates models based on transfer learning and Neural Architecture Search (NAS), selects parameters for model training, and performs model optimization. This helps service developers quickly complete model training and deployment. Based on the labeled data and application scenario provided by developers, ModelArts automatically generates models that meet precision requirements, without the need for coding. The application scenarios include image classification, object detection, and predictive analytics. Models can be automatically optimized and generated based on the deployment environment and inference speed requirements.
+ Figure 1 Process of using ExeML
+ ModelArts ExeML also provides the auto learning white-box capabilities. It opens model parameters and implements template-based development. ExeML helps accelerate the development speed. With ExeML, developers can directly optimize the generated model or retrain the model, instead of setting up a new model.
+ The key techniques of ExeML are tree search-based optimal feature transformation and Bayesian optimization for automatic parameter adjustment based on the maximum information entropy model. With these key technologies, data features and patterns can be automatically learned from enterprise relational (structured) data, and machine learning models and parameters can be intelligently optimized, achieving the same level accuracy that is achieved by experts. The key technologies of auto deep learning are transfer learning (high-quality models are generated based on a small amount of data), auto design of the model architecture in multiple dimensions (neural network search and adaptive model optimization), and auto optimization and training of training parameters.
+
+
+
+
+
\ No newline at end of file
diff --git a/docs/modelarts/umn/modelarts_01_0020.html b/docs/modelarts/umn/modelarts_01_0020.html
deleted file mode 100644
index b5e37340..00000000
--- a/docs/modelarts/umn/modelarts_01_0020.html
+++ /dev/null
@@ -1,15 +0,0 @@
-
-
- ExeML
- To implement AI in various industries, AI model development must be simplified. Currently, only a few algorithm engineers and researchers are capable of AI development and optimization. They find it challenging to develop related prototypes into products and projects. Most service developers, however, face difficulties in developing AI algorithms and optimizing parameters. As a result, most enterprises lack comprehensive AI development capabilities.
- ModelArts uses machine learning to help service developers who are not experienced in algorithm development to develop algorithms. It automatically generates models based on transfer learning and Neural Architecture Search (NAS), selects parameters for model training, and performs model optimization. This helps service developers quickly complete model training and deployment. Based on the labeled data and application scenario provided by developers, ModelArts automatically generates models that meet precision requirements, without the need for coding. The application scenarios include image classification, object detection, and predictive analytics. Models can be automatically optimized and generated based on the deployment environment and inference speed requirements.
- Figure 1 Process of using ExeML
- ModelArts ExeML also provides the auto learning white-box capabilities. It opens model parameters and implements template-based development. ExeML helps accelerate the development speed. With ExeML, developers can directly optimize the generated model or retrain the model, instead of setting up a new model.
- The key techniques of ExeML are tree search-based optimal feature transformation and Bayesian optimization for automatic parameter adjustment based on the maximum information entropy model. With these key technologies, data features and patterns can be automatically learned from enterprise relational (structured) data, and machine learning models and parameters can be intelligently optimized, achieving the same level accuracy that is achieved by experts. The key technologies of auto deep learning are transfer learning (high-quality models are generated based on a small amount of data), auto design of the model architecture in multiple dimensions (neural network search and adaptive model optimization), and auto optimization and training of training parameters.
-
-
-
diff --git a/docs/modelarts/umn/modelarts_01_0021.html b/docs/modelarts/umn/modelarts_01_0021.html
deleted file mode 100644
index e3f22302..00000000
--- a/docs/modelarts/umn/modelarts_01_0021.html
+++ /dev/null
@@ -1,39 +0,0 @@
-
-
- Billing
- ModelArts is a one-stop AI development platform geared toward developers and data scientists of all skill levels. It enables you to rapidly build, train, and deploy models anywhere (from the cloud to the edge), and manage full-lifecycle AI workflows. ModelArts accelerates AI development and fosters AI innovation with key capabilities, including data preprocessing and auto labeling, distributed training, automated model building, and one-click workflow executing.
- ModelArts is flexibly billed based on service duration.
- Billing ItemsThe fees vary based on selected ModelArts resources. For details, see Table 1.
-
- Table 1 Billing itemsBilling Item
- |
-Description
- |
-
|---|
-
-Full-process AI development
- |
-Dedicated for developers with AI development experience. ModelArts offers machine learning and deep learning algorithm development and deployment, including data processing, model development, training, and management, and service deployment. The billing items include model development environments (Notebook), model training (training and virtualization jobs), and service deployment (real-time services).
- |
-
-ExeML
- |
-Dedicated for developers with few AI development experience. ModelArts enables code-free custom model development by automatically designing, tuning, and training models based on labeled data, and deploying the models as services. Only the resources used for training and deploying ExeML jobs are billed. The billing items include training jobs, GPU deployment, and CPU deployment in ExeML.
- |
-
-
-
-
-
- Billing ModesPay-per-use: allows you to flexibly enable or disable ModelArts.
-
- Configuration Modification-related BillingYou can manually adjust the capacity of a pay-per-use dedicated resource pool. The billing is based on the number of nodes after adjustment.
- If the configuration modification methods provided by ModelArts do not meet your requirements, re-create a job and migrate data to it for configuration modification.
-
-
-
-
diff --git a/docs/modelarts/umn/modelarts_01_0023.html b/docs/modelarts/umn/modelarts_01_0023.html
new file mode 100644
index 00000000..82d08913
--- /dev/null
+++ b/docs/modelarts/umn/modelarts_01_0023.html
@@ -0,0 +1,164 @@
+
+
+ What Are the Events and Their Types for a Service?
+ During the whole lifecycle of a service, every key event is automatically recorded. You can view the events on the details page of the service at any time.
+ The following table lists the events, based on which you can locate faults occurred during service deployment and running.
+
+ Type
+ |
+Event (xxx should be replaced with the actual value.)
+ |
+
+Normal
+ |
+The service starts to deploy.
+ |
+
+Abnormal
+ |
+Insufficient resources. Wait until idle resources are sufficient.
+ |
+
+Normal
+ |
+The image starts to create.
+ |
+
+Abnormal
+ |
+Failed to create model image xxx. For details, see logs :\nxxx.
+ |
+
+Abnormal
+ |
+Failed to create the image.
+ |
+
+Normal
+ |
+The image created.
+ |
+
+Abnormal
+ |
+Service xxx failed. Error: xxx
+ |
+
+Abnormal
+ |
+Failed to update the service. Perform a rollback.
+ |
+
+Normal
+ |
+The service is being updated.
+ |
+
+Normal
+ |
+The service is being started.
+ |
+
+Normal
+ |
+The service is being stopped.
+ |
+
+Normal
+ |
+The service has been stopped.
+ |
+
+Normal
+ |
+Auto stop has been disabled.
+ |
+
+Normal
+ |
+Auto stop has been enabled. The service will stop after xs.
+ |
+
+Normal
+ |
+The service stops when the auto stop time expires.
+ |
+
+Abnormal
+ |
+The service is stopped because the quota exceeds the upper limit.
+ |
+
+Abnormal
+ |
+Failed to automatically stop the service. Error: xxx
+ |
+
+Normal
+ |
+Service instances deleted from resource pool xxx.
+ |
+
+Normal
+ |
+Service instances stopped in resource pool xxx.
+ |
+
+Abnormal
+ |
+The batch service failed. Try again later. Error: xxx
+ |
+
+Normal
+ |
+The service has been executed.
+ |
+
+Abnormal
+ |
+Failed to stop the service. Error: xxx
+ |
+
+Normal
+ |
+The subscription license xxx is to expire.
+ |
+
+Normal
+ |
+Service xxx started.
+ |
+
+Abnormal
+ |
+Failed to start service xxx.
+ |
+
+Abnormal
+ |
+Service deployment timed out. Error: xxx
+ |
+
+Normal
+ |
+Failed to update the service. The update has been rolled back.
+ |
+
+Abnormal
+ |
+Failed to update the service. The rollback failed.
+ |
+
+
+
+
+ During service deployment and running, key events can both be manually and automatically refreshed.
+ Viewing Events- In the left navigation pane of the ModelArts management console, choose Service Deployment > Real-Time Services or Batch Services or Edge Services. In the service list, click the name or ID of the target service to go to its details page.
- View the events on the Events tab page.
+
+
+
+
diff --git a/docs/modelarts/umn/modelarts_01_0028.html b/docs/modelarts/umn/modelarts_01_0028.html
new file mode 100644
index 00000000..dfa340b8
--- /dev/null
+++ b/docs/modelarts/umn/modelarts_01_0028.html
@@ -0,0 +1,39 @@
+
+
+
+ Introduction to Development Tools
+
+ This document describes the DevEnviron notebook functions of the new version.
+
+ Software development is a process of reducing developer costs and improving development experience. In AI development, ModelArts is dedicated to improving AI development experience and simplifying the development process. ModelArts DevEnviron uses cloud native resources and integrates the development tool chain to provide better in-cloud AI development experience for AI development, exploration, and teaching.
+ ModelArts notebook for seamless in-cloud and on-premises collaboration
+ - In-cloud JupyterLab, local IDE, and ModelArts plug-ins for remote development and debugging, tailored to your needs
- In-cloud development environment with AI compute resources, cloud storage, and built-in AI engines
- Custom runtime environment saved as an image for training and inference
+
+ Feature 1: Remote development, allowing remote access to notebook from a local IDEThe notebook of the new version provides remote development. After enabling remote SSH, you can remotely access the ModelArts notebook development environment to debug and run code from a local IDE.
+ Due to limited local resources, developers using a local IDE run and debug code typically on a CPU or GPU server shared between team members. Building and maintaining the CPU or GPU server are costly.
+ ModelArts notebook instances are out of the box with various built-in engines and flavors for you to select. You can use a dedicated container environment. Only after simple configurations, you can remotely access the environment to run and debug code from your local IDE.
+ ModelArts notebook can be regarded as an extension of a local development environment. The operations such as data reading, training, and file saving are the same as those performed in a local environment.
+ ModelArts notebook allows you to use in-cloud resources while with local coding habits unchanged.
+ A local IDE supports Visual Studio (VS) Code, PyCharm, and SSH. In addition, the PyCharm Toolkit plug-ins allow you to easily use cloud resources.
+
+ Feature 2: Preset images that are out-of-the-box with optimized configurations and supporting mainstream AI enginesThe AI engines and versions preset in each image are fixed. When creating a notebook instance, specify an AI engine and version, including the chip type.
+ ModelArts DevEnviron provides a group of preset images, including PyTorch, TensorFlow, and MindSpore images. You can use a preset image to start your notebook instance. After the development in the instance, submit a training job without any adaptation.
+ The image versions preset in ModelArts are determined based on user feedback and version stability. If your development can be carried out using the versions preset in ModelArts, for example, MindSpore 1.5, use preset images. These images have been fully verified and have many commonly-used installation packages built in. They are out-of-the-box, relieving you from configuring the environment.
+ The images preset in ModelArts DevEnviron include:
+
+ - To simplify operations, ModelArts notebook of the new version does not support switchover between AI engines in a notebook instance.
- AI engines vary based on regions. For details about the AI engines available in a region, see the AI engines displayed on the management console.
+
+
+ Feature 3: JupyterLab, an online interactive development and debugging toolModelArts integrates open-source JupyterLab for online interactive development and debugging. You can use the notebook on the ModelArts management console to compile and debug code and train models based on the code, without concerning environment installation or configuration.
+ JupyterLab is an interactive development environment. It is the next-generation product of Jupyter Notebook. JupyterLab enables you to compile notebooks, operate terminals, edit Markdown text, enable interaction, and view CSV files and images.
+
+
+
+
+
diff --git a/docs/modelarts/umn/modelarts_04_0099.html b/docs/modelarts/umn/modelarts_04_0099.html
deleted file mode 100644
index 9b992315..00000000
--- a/docs/modelarts/umn/modelarts_04_0099.html
+++ /dev/null
@@ -1,20 +0,0 @@
-
-
- Change History
-
- Released On
- |
-Description
- |
-
|---|
-
-2024-03-30
- |
-This is the first official release.
- |
-
-
-
-
-
-
diff --git a/docs/modelarts/umn/modelarts_05_0001.html b/docs/modelarts/umn/modelarts_05_0001.html
index edeef704..2bbd9540 100644
--- a/docs/modelarts/umn/modelarts_05_0001.html
+++ b/docs/modelarts/umn/modelarts_05_0001.html
@@ -1,10 +1,10 @@
-
+
What Is ModelArts?
- ModelArts is a one-stop AI development platform geared toward developers and data scientists of all skill levels. It enables you to rapidly build, train, and deploy models anywhere (from the cloud to the edge), and manage full-lifecycle AI workflows. ModelArts accelerates AI development and fosters AI innovation with key capabilities, including data preprocessing and auto labeling, distributed training, automated model building, and one-click workflow executing.
- The one-stop ModelArts platform covers all stages of AI development, including data processing, AI application creation, and model training and deployment. The underlying layer of ModelArts supports various heterogeneous computing resources. You can flexibly select and use the resources without having to consider the underlying technologies. In addition, ModelArts supports popular open-source AI development frameworks such as TensorFlow. Developers can also use self-developed algorithm frameworks to match their usage habits.
-
- ModelArts aims to achieve simple, convenient AI development.
+ ModelArts is a one-stop AI development platform geared toward developers and data scientists of all skill levels. It enables you to rapidly build, train, and deploy models anywhere (from the cloud to the edge), and manage full-lifecycle AI workflows. ModelArts accelerates AI development and fosters AI innovation with key capabilities, including data preprocessing and auto labeling, distributed training, automated model building, and one-click workflow executing.
+ The one-stop ModelArts platform covers all stages of AI development, including data processing, AI application creation, and model training and deployment. The underlying layer of ModelArts supports various heterogeneous computing resources. You can flexibly select and use the resources without having to consider the underlying technologies. In addition, ModelArts supports popular open-source AI development frameworks such as TensorFlow. Developers can also use self-developed algorithm frameworks to match their usage habits.
+
+ ModelArts aims to achieve simple, convenient AI development.
diff --git a/docs/modelarts/umn/modelarts_05_0002.html b/docs/modelarts/umn/modelarts_05_0002.html
deleted file mode 100644
index 950eef0c..00000000
--- a/docs/modelarts/umn/modelarts_05_0002.html
+++ /dev/null
@@ -1,12 +0,0 @@
-
-
- What Is ExeML?
- ExeML is the process of automating model design, parameter tuning, and model training, compression, and deployment with the labeled data. The process is free of coding and does not require developers' experience in model development.
- Users who do not have encoding capability can use the labeling, one-click model training, and model deployment functions of ExeML to build AI models.
-
-
-
diff --git a/docs/modelarts/umn/modelarts_05_0003.html b/docs/modelarts/umn/modelarts_05_0003.html
index 264c2eb0..447b285a 100644
--- a/docs/modelarts/umn/modelarts_05_0003.html
+++ b/docs/modelarts/umn/modelarts_05_0003.html
@@ -1,15 +1,15 @@
-
+
What Are the Relationships Between ModelArts and Other Services?
- OBSModelArts uses Object Storage Service (OBS) to securely and reliably store data and models at low costs. For more details, see Object Storage Service Console Operation Guide.
+ OBSModelArts uses Object Storage Service (OBS) to securely and reliably store data and models at low costs. For more details, see Object Storage Service Console Operation Guide.
- CCEModelArts uses Cloud Container Engine (CCE) to deploy models as real-time services. CCE enables high concurrency and provides elastic scaling. For more information about CCE, see Cloud Container Engine User Guide.
+ CCEModelArts uses Cloud Container Engine (CCE) to deploy models as real-time services. CCE enables high concurrency and provides elastic scaling. For more information about CCE, see Cloud Container Engine User Guide.
- SWRTo use an AI framework that is not supported by ModelArts, use Software Repository for Container (SWR) to customize an image and import the image to ModelArts for training or inference. For details about SWR, see .
+ SWRTo use an AI framework that is not supported by ModelArts, use Software Repository for Container (SWR) to customize an image and import the image to ModelArts for training or inference. For details about SWR, see .
- Cloud EyeModelArts uses Cloud Eye to monitor online services and model loads in real time and send alarms and notifications automatically. For details about Cloud Eye, see Cloud Eye User Guide.
+ Cloud EyeModelArts uses Cloud Eye to monitor online services and model loads in real time and send alarms and notifications automatically. For details about Cloud Eye, see Cloud Eye User Guide.
- CTSModelArts uses Cloud Trace Service (CTS) to record operations for later query, audit, and backtrack operations. For details about CTS, see Cloud Trace Service User Guide.
+ CTSModelArts uses Cloud Trace Service (CTS) to record operations for later query, audit, and backtrack operations. For details about CTS, see Cloud Trace Service User Guide.
diff --git a/docs/modelarts/umn/modelarts_05_0004.html b/docs/modelarts/umn/modelarts_05_0004.html
index b3a7d0ed..125ba8a1 100644
--- a/docs/modelarts/umn/modelarts_05_0004.html
+++ b/docs/modelarts/umn/modelarts_05_0004.html
@@ -1,7 +1,7 @@
-
+
How Do I Obtain an Access Key?
- Obtaining an Access Key- Log in to the console, enter the My Credentials page, and choose Access Keys > Create Access Key.
- In the Create Access Key dialog box that is displayed, use the login password for verification.
- Click OK, open the credentials.csv file, and save the key file as prompted. The access key file is saved in the default downloads folder of the browser. Then, the access key (Access Key Id and Secret Access Key) is obtained.
+ Obtaining an Access Key- Log in to the console, enter the My Credentials page, and choose Access Keys > Create Access Key.
- In the Create Access Key dialog box that is displayed, use the login password for verification.
- Click OK, open the credentials.csv file, and save the key file as prompted. The access key file is saved in the default downloads folder of the browser. Then, the access key (Access Key Id and Secret Access Key) is obtained.
diff --git a/docs/modelarts/umn/modelarts_05_0005.html b/docs/modelarts/umn/modelarts_05_0005.html
deleted file mode 100644
index 7da016a2..00000000
--- a/docs/modelarts/umn/modelarts_05_0005.html
+++ /dev/null
@@ -1,11 +0,0 @@
-
-
- What Should I Do When the Train Button Is Unavailable After I Create an Image Classification Project and Label the Images?
- The Train button turns to be available when the training images for an image classification project are classified into at least two categories, and each category contains at least five images.
-
-
-
diff --git a/docs/modelarts/umn/modelarts_05_0006.html b/docs/modelarts/umn/modelarts_05_0006.html
deleted file mode 100644
index e09abd02..00000000
--- a/docs/modelarts/umn/modelarts_05_0006.html
+++ /dev/null
@@ -1,11 +0,0 @@
-
-
- Can I Add Multiple Labels to an Image for an Object Detection Project?
- Yes. You can add multiple labels to an image.
-
-
-
diff --git a/docs/modelarts/umn/modelarts_05_0008.html b/docs/modelarts/umn/modelarts_05_0008.html
deleted file mode 100644
index 9abb10fa..00000000
--- a/docs/modelarts/umn/modelarts_05_0008.html
+++ /dev/null
@@ -1,12 +0,0 @@
-
-
- What Type of Service Is Deployed in ExeML?
- Models created in ExeML are deployed as real-time services. You can add images or compile code to test the services, as well as call the APIs using the URLs.
- After model development is successful, you can choose Service Deployment > Real-Time Services in the left navigation pane of the ModelArts console to view running services, and stop or delete services.
-
-
-
diff --git a/docs/modelarts/umn/modelarts_05_0010.html b/docs/modelarts/umn/modelarts_05_0010.html
deleted file mode 100644
index d7c4df36..00000000
--- a/docs/modelarts/umn/modelarts_05_0010.html
+++ /dev/null
@@ -1,11 +0,0 @@
-
-
- What Formats of Images Are Supported by Object Detection or Image Classification Projects?
- Images in JPG, JPEG, PNG, or BMP format are supported.
-
-
-
diff --git a/docs/modelarts/umn/modelarts_05_0012.html b/docs/modelarts/umn/modelarts_05_0012.html
index b259f8b7..74531a4b 100644
--- a/docs/modelarts/umn/modelarts_05_0012.html
+++ b/docs/modelarts/umn/modelarts_05_0012.html
@@ -1,11 +1,11 @@
-
+
What Types of Services Can Models Be Deployed as on ModelArts?
- Models can be deployed as real-time services or batch services.
+ Models can be deployed as real-time services or batch services.
diff --git a/docs/modelarts/umn/modelarts_05_0013.html b/docs/modelarts/umn/modelarts_05_0013.html
index a0f349b2..f2a367af 100644
--- a/docs/modelarts/umn/modelarts_05_0013.html
+++ b/docs/modelarts/umn/modelarts_05_0013.html
@@ -1,7 +1,7 @@
-
+
How Do I Upload Data to OBS?
- Before using ModelArts to develop AI models, data needs to be uploaded to an OBS bucket. You can log in to the OBS console to create an OBS bucket, create a folder in it, and upload data. For details about how to upload data, see Object Storage Service User Guide.
+ Before using ModelArts to develop AI models, data needs to be uploaded to an OBS bucket. You can log in to the OBS console to create an OBS bucket, create a folder in it, and upload data. For details about how to upload data, see Object Storage Service User Guide.
diff --git a/docs/modelarts/umn/modelarts_05_0014.html b/docs/modelarts/umn/modelarts_05_0014.html
index 086e5e6d..34ed8a13 100644
--- a/docs/modelarts/umn/modelarts_05_0014.html
+++ b/docs/modelarts/umn/modelarts_05_0014.html
@@ -1,23 +1,41 @@
-
+
General Issues
-
+
diff --git a/docs/modelarts/umn/modelarts_05_0015.html b/docs/modelarts/umn/modelarts_05_0015.html
deleted file mode 100644
index 46113b55..00000000
--- a/docs/modelarts/umn/modelarts_05_0015.html
+++ /dev/null
@@ -1,29 +0,0 @@
-
-
- ExeML
-
-
-
diff --git a/docs/modelarts/umn/modelarts_05_0016.html b/docs/modelarts/umn/modelarts_05_0016.html
index 94805dbb..c2e39ee5 100644
--- a/docs/modelarts/umn/modelarts_05_0016.html
+++ b/docs/modelarts/umn/modelarts_05_0016.html
@@ -1,17 +1,15 @@
-
+
Model Management
-
+
diff --git a/docs/modelarts/umn/modelarts_05_0017.html b/docs/modelarts/umn/modelarts_05_0017.html
index d912b049..47f5557f 100644
--- a/docs/modelarts/umn/modelarts_05_0017.html
+++ b/docs/modelarts/umn/modelarts_05_0017.html
@@ -1,17 +1,17 @@
-
+
Service Deployment
-
+
diff --git a/docs/modelarts/umn/modelarts_05_0018.html b/docs/modelarts/umn/modelarts_05_0018.html
deleted file mode 100644
index 4ffbb73a..00000000
--- a/docs/modelarts/umn/modelarts_05_0018.html
+++ /dev/null
@@ -1,14 +0,0 @@
-
-
- What Are Image Classification and Object Detection?
- Image classification is an image processing method that separates different classes of targets according to the features reflected in the images. With quantitative analysis on images, it classifies an image or each pixel or area in an image into different categories to replace human visual interpretation. In general, image classification aims to identify a class, status, or scene in an image. It is applicable to scenarios where an image contains only one object. The following figure shows an example of identifying a car in an image.
- Figure 1 Image classification
- Object detection is one of the classical problems in computer vision. It intends to label objects with frames and identify the object classes in an image. Generally, if an image contains multiple objects, object detection can identify the location, quantity, and name of each object in the image. It is suitable for scenarios where an image contains multiple objects. Figure 2 shows an example of identifying a tree and a car in an image.
- Figure 2 Object detection
-
-
-
diff --git a/docs/modelarts/umn/modelarts_05_0019.html b/docs/modelarts/umn/modelarts_05_0019.html
new file mode 100644
index 00000000..5a674394
--- /dev/null
+++ b/docs/modelarts/umn/modelarts_05_0019.html
@@ -0,0 +1,16 @@
+
+
+ What Do I Do If the System Displays a Message Indicating that the AK/SK Pair Is Unavailable?
+ Issue AnalysisAn AK and SK form a key pair required for accessing OBS. Each SK corresponds to a specific AK, and each AK corresponds to a specific user. If the system displays a message indicating that the AK/SK pair is unavailable, it is possible that the account is in arrears or the AK/SK pair is incorrect.
+
+ Solution- Use the current account to log in to the OBS console and check whether the current account can access OBS.
- If the account can access OBS, rectify the fault by referring to 2.
+ - If yes, .
- If not, replace the AK/SK with those created using the current account. For details, see Access Keys.
+
+
+
+
+
diff --git a/docs/modelarts/umn/modelarts_05_0020.html b/docs/modelarts/umn/modelarts_05_0020.html
new file mode 100644
index 00000000..ff3b589b
--- /dev/null
+++ b/docs/modelarts/umn/modelarts_05_0020.html
@@ -0,0 +1,29 @@
+
+
+ Notebook (New Version)
+
+
+
diff --git a/docs/modelarts/umn/modelarts_05_0021.html b/docs/modelarts/umn/modelarts_05_0021.html
new file mode 100644
index 00000000..0ac7458f
--- /dev/null
+++ b/docs/modelarts/umn/modelarts_05_0021.html
@@ -0,0 +1,24 @@
+
+
+ What Do I Do If I Cannot Access My Notebook Instance?
+ Troubleshoot the issue based on error code.
+ Error 404If this error is reported when an IAM user creates an instance, the IAM user does not have the permissions to access the corresponding storage location (OBS bucket).
+ Solution
+ - Log in to the OBS console using the primary account and grant access permissions for the OBS bucket to the IAM user.
- After the IAM user obtains the permissions, log in to the ModelArts console, delete the instance, and use the OBS path to create a notebook instance.
+
+ Error 503If this error is reported, it is possible that the instance is consuming too many resources. If this is the case, stop the instance and restart it.
+
+ Error 500Notebook JupyterLab cannot be opened, and error 500 is reported. The possible cause is that the disk space in the work directory is used up. In this case, identify the fault cause and clear the disk by referring to .
+
+ Error "This site can't be reached"After a notebook instance is created, click Open in the Operation column. The error message shown in the following figure is displayed.
+ 
+ Do as follows to resolve this issue: Copy the domain name of the page , add it to the Do not use proxy server for addresses beginning with text box, and save the settings.
+
+
+
+
+
diff --git a/docs/modelarts/umn/modelarts_05_0022.html b/docs/modelarts/umn/modelarts_05_0022.html
new file mode 100644
index 00000000..4ee05720
--- /dev/null
+++ b/docs/modelarts/umn/modelarts_05_0022.html
@@ -0,0 +1,31 @@
+
+
+ How Do I Install External Libraries in a Notebook Instance?
+ Multiple environments such as Jupyter and Python have been integrated into ModelArts notebook to support many frameworks, including TensorFlow, MindSpore, PyTorch, and Spark. You can use pip install to install external libraries in Jupyter Notebook or on the Terminal page.
+ Installing External Libraries in Jupyter NotebookYou can use JupyterLab to install Shapely in the TensorFlow-1.8 environment.
+ - Open a notebook instance and access the Launcher page.
- In the Notebook area, click TensorFlow-1.8 and create an IPYNB file.
- In the new notebook instance, enter the following command in the code input bar:
!pip install Shapely
+
+
+ Installing External Libraries on the Terminal PageYou can use pip to install external libraries in the TensorFlow-1.8 environment on the Terminal page. For example, to install Shapely:
+ - Open a notebook instance and access the Launcher page.
- In the Other area, click Terminal and create a terminal file.
- Enter the following commands in the code input box to obtain the kernel of the current environment and activate the Python environment on which the installation depends:
cat /home/ma-user/README
+source /home/ma-user/anaconda3/bin/activate TensorFlow-1.8
+ To install TensorFlow in another Python environment, replace TensorFlow-1.8 in the command with the target engine.
+
+Figure 1 Activating the environment
+ - Run the following command in the code input box to install Shapely:
pip install Shapely
+
+
+
+
+
+
+
\ No newline at end of file
diff --git a/docs/modelarts/umn/modelarts_05_0024.html b/docs/modelarts/umn/modelarts_05_0024.html
index 3809e9d8..46ed7e6c 100644
--- a/docs/modelarts/umn/modelarts_05_0024.html
+++ b/docs/modelarts/umn/modelarts_05_0024.html
@@ -1,26 +1,26 @@
-
+
How Do I Upload a File from a Notebook Instance to OBS or Download a File from OBS to a Notebook Instance?
- In a notebook instance, you can call the ModelArts MoXing API or SDK to exchange data with OBS for uploading a file to OBS or downloading a file from OBS to the notebook instance.
- Figure 1 Uploading or downloading a file
- Method 1: Using MoXing to Upload and Download a FileDeveloped by the ModelArts team, MoXing is a distributed training acceleration framework built on open-source deep learning engines such as TensorFlow and PyTorch. MoXing makes model coding easier and more efficient.
- MoXing provides a set of file object APIs for reading and writing OBS files.
- Sample code:
- import moxing as mox
+In a notebook instance, you can call the ModelArts MoXing API or SDK to exchange data with OBS for uploading a file to OBS or downloading a file from OBS to the notebook instance.
+ Figure 1 Uploading or downloading a file
+ Method 1: Using MoXing to Upload and Download a FileDeveloped by the ModelArts team, MoXing is a distributed training acceleration framework built on open-source deep learning engines such as TensorFlow and PyTorch. MoXing makes model coding easier and more efficient.
+ MoXing provides a set of file object APIs for reading and writing OBS files.
+ Sample code:
+ import moxing as mox
-# Download the OBS folder sub_dir_0 from OBS to a notebook instance.
+# Download the OBS folder sub_dir_0 from OBS to a notebook instance.
mox.file.copy_parallel('obs://bucket_name/sub_dir_0', '/home/ma-user/work/sub_dir_0')
-# Download the OBS file obs_file.txt from OBS to a notebook instance.
+# Download the OBS file obs_file.txt from OBS to a notebook instance.
mox.file.copy('obs://bucket_name/obs_file.txt', '/home/ma-user/work/obs_file.txt')
-# Upload the OBS folder sub_dir_0 from the notebook instance to OBS.
+# Upload the OBS folder sub_dir_0 from a notebook instance to OBS.
mox.file.copy_parallel('/home/ma-user/work/sub_dir_0', 'obs://bucket_name/sub_dir_0')
-# Upload the OBS file obs_file.txt from a notebook instance to OBS.
+# Upload the OBS file obs_file.txt from a notebook instance to OBS.
mox.file.copy('/home/ma-user/work/obs_file.txt', 'obs://bucket_name/obs_file.txt')
- Method 2: Use SDKs to Exchange DataCall the ModelArts SDK for downloading a file from OBS.
- Sample code: Download file1.txt from OBS to /home/ma-user/work/ in the notebook instance. All the bucket name, folder name, and file name are customizable.
- 1
+Method 2: Using SDK to Upload and Download a FileCall the ModelArts SDK for downloading a file from OBS.
+ Sample code: Download file1.txt from OBS to /home/ma-user/work/ in the notebook instance. All the bucket name, folder name, and file name are customizable.
+ | from modelarts.session import Session
session = Session()
@@ -28,14 +28,14 @@ mox.file.copy('/home/ma-user/work/obs_file.txt', 'obs://bucket_name/obs_file.txt
|
- Call the ModelArts SDK for downloading a folder from OBS.
- Sample code: Download dir1 from OBS to /home/ma-user/work/ in the notebook instance. The bucket name and folder name are customizable.
- from modelarts.session import Session
+Call the ModelArts SDK for downloading a folder from OBS.
+Sample code: Download dir1 from OBS to /home/ma-user/work/ in the notebook instance. The bucket name and folder name are customizable.
+from modelarts.session import Session
session = Session()
session.obs.download_dir(src_obs_dir="obs://bucket-name/dir1/", dst_local_dir="/home/ma-user/work/")
-Call the ModelArts SDK for uploading a file to OBS.
-Sample code: Upload file1.txt in the notebook instance to OBS bucket obs://bucket-name/dir1/. All the bucket name, folder name, and file name are customizable.
-1
+Call the ModelArts SDK for uploading a file to OBS.
+Sample code: Upload file1.txt in the notebook instance to OBS bucket obs://bucket-name/dir1/. All the bucket name, folder name, and file name are customizable.
+ | from modelarts.session import Session
session = Session()
@@ -43,16 +43,23 @@ session.obs.download_dir(src_obs_dir="obs://bucket-name/dir1/", dst_local_dir="/
|
Call the ModelArts SDK for uploading a folder to OBS.
-Sample code: Upload /work/ in the notebook instance to obs://bucket-name/dir1/work/ of bucket-name. The bucket name and folder name are customizable.
-from modelarts.session import Session
+Call the ModelArts SDK for uploading a folder to OBS.
+Sample code: Upload /work/ in the notebook instance to obs://bucket-name/dir1/work/ of bucket-name. The bucket name and folder name are customizable.
+from modelarts.session import Session
session = Session()
session.obs.upload_dir(src_local_dir='/home/ma-user/work/', dst_obs_dir='obs://bucket-name/dir1/')
Training Jobs
-
+
diff --git a/docs/modelarts/umn/modelarts_05_0031.html b/docs/modelarts/umn/modelarts_05_0031.html
index 3e7e27f5..4f25997a 100644
--- a/docs/modelarts/umn/modelarts_05_0031.html
+++ b/docs/modelarts/umn/modelarts_05_0031.html
@@ -1,14 +1,14 @@
-
+
What Can I Do If the Message "Object directory size/quantity exceeds the limit" Is Displayed When I Create a Training Job?
-Issue AnalysisThe code directory for creating a training job has limits on the size and number of files.
+ Issue AnalysisThe code directory for creating a training job has limits on the size and number of files.
- SolutionDelete the files except the code from the code directory or save the files in other directories. Ensure that the size of the code directory does not exceed 128 MB and the number of files does not exceed 4,096.
+ SolutionDelete the files except the code from the code directory or save the files in other directories. Ensure that the size of the code directory does not exceed 128 MB and the number of files does not exceed 4,096.
diff --git a/docs/modelarts/umn/modelarts_05_0032.html b/docs/modelarts/umn/modelarts_05_0032.html
deleted file mode 100644
index eb2c2c8d..00000000
--- a/docs/modelarts/umn/modelarts_05_0032.html
+++ /dev/null
@@ -1,34 +0,0 @@
-
-
- What Do I Do If Error Message "No such file or directory" Is Displayed in Training Job Logs?
- When you use ModelArts, your data is stored in an OBS bucket. There is an OBS path to your data, for example, bucket_name/dir/image.jpg. ModelArts training jobs run in containers, and the jobs access OBS data through the OBS path to the data. If the file or path is unavailable, it is possible that the data storage path was incorrectly configured when you created the training job, or that the access path in the code file is incorrect.
- Perform the following operations to locate the fault:
- - Checking Whether the Affected Path Is an OBS Path
- Checking Whether the Affected Path Is Available
- Checking Whether the Affected Path Is an OBS PathWhen using ModelArts, store your data in an OBS bucket. However, the OBS path cannot be used to read data during the execution of the training code.
- The reason is as follows:
- After a training job is created, the training performance is poor if the running container is directly connected to OBS. To prevent this issue, the system automatically downloads the training data to the local path of the running container. Therefore, an error occurs if an OBS path is used in training code.
- If the affected path is a path to the training data, perform the following operations to resolve this issue (see input and output configurations for details):
- - When creating an algorithm, set the code path parameter, which defaults to data_url, in the input path mapping configuration.
- Add a hyperparameter, which defaults to data_url, to the training code. Use data_url as the local path for inputting the training data.
-
- Checking Whether the Affected Path Is AvailableThe code developed locally must be uploaded to the ModelArts backend. In training code, it is error-prone to set the path for storing the dependency file.
- The following general solution is recommended: Use the OS API to obtain the absolute path of the dependency file.
- Example:
- |---project_root # Root directory for code
- |---BootfileDirectory # Directory where the boot file is located
- |---bootfile.py # Boot file
- |---otherfileDirectory # Directory of other dependency files
- |---otherfile.py # Other dependency files
-
- Do as follows to obtain the path of the dependency file, otherfile_path in this example, in the boot file:
- import os
-current_path = os.path.dirname(os.path.realpath(__file__)) # Directory where the boot file is located
-project_root = os.path.dirname(current_path) # Root directory of the project, which is the code directory set on the ModelArts training console
-otherfile_path = os.path.join(project_root, "otherfileDirectory", "otherfile.py")
-
-
-
-
diff --git a/docs/modelarts/umn/modelarts_05_0041.html b/docs/modelarts/umn/modelarts_05_0041.html
new file mode 100644
index 00000000..cf5e829a
--- /dev/null
+++ b/docs/modelarts/umn/modelarts_05_0041.html
@@ -0,0 +1,11 @@
+
+
+ How Do I Use ModelArts to Train Models Based on Structured Data?
+ For more advanced users, ModelArts provides the notebook creation function of DevEnviron for code development. It allows the users to create training tasks with large volumes of data in training jobs and use the engines such as Scikit_Learn, XGBoost, or Spark_MLlib in the development and training processes.
+
+
+
diff --git a/docs/modelarts/umn/modelarts_05_0042.html b/docs/modelarts/umn/modelarts_05_0042.html
new file mode 100644
index 00000000..618f4130
--- /dev/null
+++ b/docs/modelarts/umn/modelarts_05_0042.html
@@ -0,0 +1,15 @@
+
+
+ Is the Keras Engine Supported?
+ Notebook instances in DevEnviron support the Keras engine. The Keras engine is not supported in job training and model deployment (inference).
+ Keras is an advanced neural network API written in Python. It is capable of running on top of TensorFlow, CNTK, or Theano. Notebook instances in DevEnviron support tf.keras.
+ How Do I View Keras Versions?- On the ModelArts management console, create a notebook instance with image TensorFlow-1.13 or TensorFlow-1.15.
- Access the notebook instance. In JupyterLab, run !pip list to view Keras versions.
Figure 1 Viewing Keras versions
+
+
+
+
+
diff --git a/docs/modelarts/umn/modelarts_05_0044.html b/docs/modelarts/umn/modelarts_05_0044.html
new file mode 100644
index 00000000..01d2ce3a
--- /dev/null
+++ b/docs/modelarts/umn/modelarts_05_0044.html
@@ -0,0 +1,14 @@
+
+
+ What Should I Do When the System Displays an Error Message Indicating that No Space Left After I Run the pip install Command?
+ SymptomIn the notebook instance, error message "No Space left..." is displayed after the pip install command is run.
+
+ SolutionYou are advised to run the pip install --no-cache ** command instead of the pip install ** command. Adding the --no-cache parameter can solve such problem.
+
+
+
+
diff --git a/docs/modelarts/umn/modelarts_05_0045.html b/docs/modelarts/umn/modelarts_05_0045.html
index b5d7a61e..0923daec 100644
--- a/docs/modelarts/umn/modelarts_05_0045.html
+++ b/docs/modelarts/umn/modelarts_05_0045.html
@@ -1,11 +1,11 @@
-
+
Where Will the Data Be Uploaded to?
- If you use OBS to store the notebook instance, after you click upload, the data is directly uploaded to the target OBS path, that is, the OBS path specified when the notebook instance is created.
+ If you use OBS to store the notebook instance, after you click upload, the data is directly uploaded to the target OBS path, that is, the OBS path specified when the notebook instance is created.
diff --git a/docs/modelarts/umn/modelarts_05_0050.html b/docs/modelarts/umn/modelarts_05_0050.html
new file mode 100644
index 00000000..75d2b108
--- /dev/null
+++ b/docs/modelarts/umn/modelarts_05_0050.html
@@ -0,0 +1,12 @@
+
+
+ Why Does the Instance Break Down When dead kernel Is Displayed During Training Code Running?
+ The notebook instance breaks down during training code running due to insufficient memory caused by large data volume or excessive training layers.
+ After this error occurs, the system automatically restarts the notebook instance to fix the instance breakdown. In this case, only the breakdown is fixed. If you run the training code again, the failure will still occur. To solve the problem of insufficient memory, you are advised to create a new notebook instance and use a resource pool of higher specifications, such as a GPU or dedicated resource pool, to run the training code. An existing notebook instance that has been successfully created cannot be scaled up using resources with higher specifications.
+
+
+
diff --git a/docs/modelarts/umn/modelarts_05_0051.html b/docs/modelarts/umn/modelarts_05_0051.html
new file mode 100644
index 00000000..d8b32875
--- /dev/null
+++ b/docs/modelarts/umn/modelarts_05_0051.html
@@ -0,0 +1,12 @@
+
+
+ What Do I Do If the Code Can Be Run But Cannot Be Saved, and the Error Message "save error" Is Displayed?
+ If the notebook instance can run the code but cannot save it, the error message "save error" is displayed when you save the file. In most cases, this error is caused by a security policy of Web Application Firewall (WAF).
+ On the current page, some characters in your input or output of the code are intercepted because they are considered to be a security risk. Submit a service ticket and contact customer service to check and handle the problem.
+
+
+
diff --git a/docs/modelarts/umn/modelarts_05_0052.html b/docs/modelarts/umn/modelarts_05_0052.html
new file mode 100644
index 00000000..16bff6f1
--- /dev/null
+++ b/docs/modelarts/umn/modelarts_05_0052.html
@@ -0,0 +1,12 @@
+
+
+ What Are the Differences Between ModelArts and DLS?
+ Deep Learning Service (DLS) is a one-stop deep learning platform . With various optimized neural network models, DLS allows you to easily implement model training and evaluation with the flexibility of on-demand scheduling.
+ However, DLS supports only the deep learning technologies, while ModelArts integrates both the deep learning and machine learning technologies. In addition, ModelArts is a one-stop AI development platform, which manages the AI development lifecycle from data labeling, algorithm development, to model training and deployment. To be specific, ModelArts contains and supports the functions and features of DLS. Currently, DLS is terminated . The functions related to deep learning can be directly used in ModelArts. If you are a DLS user, you can also migrate the data in DLS to ModelArts.
+
+
+
diff --git a/docs/modelarts/umn/modelarts_05_0057.html b/docs/modelarts/umn/modelarts_05_0057.html
new file mode 100644
index 00000000..c0da5641
--- /dev/null
+++ b/docs/modelarts/umn/modelarts_05_0057.html
@@ -0,0 +1,16 @@
+
+
+ How Do I Import Large Files to a Notebook Instance?
+ - Large files (files larger than 100 MB)
Use OBS to upload large files. To do so, use OBS Browser to upload a local file to an OBS bucket and use ModelArts SDK to download the file from OBS to a notebook instance.
+For details about how to use ModelArts SDK or MoXing to download files from OBS, see How Do I Upload a File from a Notebook Instance to OBS or Download a File from OBS to a Notebook Instance?
+ - Folders
Compress a folder into a package and upload the package in the same way as uploading a large file. After the package is uploaded to a notebook instance, decompress it on the Terminal page.
+unzip xxx.zip # Directly decompress the package in the path where the package is stored.
+For more details, search for the decompression command in mainstream search engines.
+
+
+
+
diff --git a/docs/modelarts/umn/modelarts_05_0058.html b/docs/modelarts/umn/modelarts_05_0058.html
new file mode 100644
index 00000000..988c056b
--- /dev/null
+++ b/docs/modelarts/umn/modelarts_05_0058.html
@@ -0,0 +1,11 @@
+
+
+ Can I Install MoXing in a Local Environment?
+ No. MoXing can be used only on ModelArts.
+
+
+
diff --git a/docs/modelarts/umn/modelarts_05_0059.html b/docs/modelarts/umn/modelarts_05_0059.html
new file mode 100644
index 00000000..66c8d192
--- /dev/null
+++ b/docs/modelarts/umn/modelarts_05_0059.html
@@ -0,0 +1,13 @@
+
+
+ What Do I Do If a Notebook Instance Won't Run My Code?
+ If a notebook instance fails to execute code, you can locate and rectify the fault as follows:
+ - If the execution of a cell is suspended or lasts for a long time (for example, the execution of the second and third cells in Figure 1 is suspended or lasts for a long time, causing execution failure of the fourth cell) but the notebook page still responds and other cells can be selected, click interrupt the kernel highlighted in a red box in the following figure to stop the execution of all cells. The notebook instance retains all variable spaces.
Figure 1 Stopping all cells
+ - If the notebook page does not respond, close the notebook page and the ModelArts console. Then, open the ModelArts console and access the notebook instance again. The notebook instance retains all the variable spaces that exist when the notebook instance is unavailable.
- If the notebook instance still cannot be used, access the Notebook page on the ModelArts console and stop the notebook instance. After the notebook instance is stopped, click Start to restart the notebook instance and open it. The instance will have preserved all the spaces for the variables that were unable to run.
+
+
+
diff --git a/docs/modelarts/umn/modelarts_05_0063.html b/docs/modelarts/umn/modelarts_05_0063.html
deleted file mode 100644
index 6e82da72..00000000
--- a/docs/modelarts/umn/modelarts_05_0063.html
+++ /dev/null
@@ -1,32 +0,0 @@
-
-
- How Do I Create a Training Job When a Dependency Package Is Referenced in a Model?
- When a model references a dependency package, select a frequently-used framework to create training jobs. In addition, place the required file or installation package in the code directory. The requirements vary based on the dependency package that you use.
- - Open-source installation package
It is not allowed to install the package using the GitHub source code.
-
-Create a file named pip-requirements.txt in the code directory. In this file, specify the name and version of the dependency package in the format of Package name==Version.
-For example, the OBS path specified by Code Directory contains model files and the pip-requirements.txt file. The following shows the code directory structure:
-|---OBS path to the model boot file
- |---model.py #Model boot file
- |---pip-requirements.txt #Customized configuration file, which specifies the name and version of the dependency package
-The following shows the content of the pip-requirements.txt file:
-alembic==0.8.6
-bleach==1.4.3
-click==6.6
- - Customized WHL file
When you use a customized .whl file, the system cannot automatically download and install the file. Place the .whl file in the code directory, create a file named pip-requirements.txt, and specify the name of the .whl file in the created file. The dependency package must be a .whl file.
-For example, the OBS path specified by Code Directory contains model files, .whl file, and pip-requirements.txt file. The following shows the code directory structure:
-|---OBS path to the model boot file
- |---model.py #Model boot file
- |---XXX.whl #Dependency package. If multiple dependencies are required, place all of them here.
- |---pip-requirements.txt #Customized configuration file, which specifies the name of the dependency package
-The following shows the content of the pip-requirements.txt file:
-numpy-1.15.4-cp36-cp36m-manylinux1_x86_64.whl
-tensorflow-1.8.0-cp36-cp36m-manylinux1_x86_64.whl
-
-
-
-
diff --git a/docs/modelarts/umn/modelarts_05_0067.html b/docs/modelarts/umn/modelarts_05_0067.html
index b9102479..7aa90344 100644
--- a/docs/modelarts/umn/modelarts_05_0067.html
+++ b/docs/modelarts/umn/modelarts_05_0067.html
@@ -1,21 +1,27 @@
-
+
- Notebook
-
+ Code Execution
+
diff --git a/docs/modelarts/umn/modelarts_05_0071.html b/docs/modelarts/umn/modelarts_05_0071.html
index 55cfa4ae..66f1e4db 100644
--- a/docs/modelarts/umn/modelarts_05_0071.html
+++ b/docs/modelarts/umn/modelarts_05_0071.html
@@ -1,12 +1,19 @@
-
+
How Do I Enable the Terminal Function in DevEnviron of ModelArts?
- - Log in to the ModelArts management console, and choose .
- In the notebook list, click Open in the Operation column of the target notebook instance to go to the Jupyter page.
- On the Files tab page of the Jupyter page, click New and select Terminal. The Terminal page is displayed.
Figure 1 Going to the Terminal page
+- Log in to the ModelArts management console, and choose .
- Create a notebook instance. When the instance is running, click Open in the Operation column. The JupyterLab page is displayed.
- Choose File > New > Terminal. The Terminal page is displayed.
Figure 1 Going to the Terminal page
How Do I Check Whether ModelArts and an OBS Bucket Are in the Same Region?
+If an OBS directory needs to be specified for using ModelArts functions, such as creating training jobs and datasets, ensure that the OBS bucket and ModelArts are in the same region.
+ Checking Whether the OBS Bucket and ModelArts Are in the Same Region- Check the region where the created OBS bucket resides.
- Log in to .
- On the Object Storage Service page, to search for a bucket, enter a keyword in Bucket Name.
In the Region column, view the region where the created OBS bucket is located.
+
+ - Check the region where ModelArts is deployed.
Log in to the ModelArts console and view the region where ModelArts resides in the upper left corner.
+ - Check whether the region of the created OBS bucket is the same as that of ModelArts. Ensure that the OBS bucket is in the same region as ModelArts.
+
+ How Do I Perform Incremental Training When Using MoXing?
+If you are not satisfied with training results when using MoXing to build a model, you can perform incremental training after modifying some data and label information.
+ Adding Incremental Training Parameters to mox.runAfter modifying labeling data or datasets, you can modify the log_dir parameter in and add the checkpoint_path parameter to mox.run. Set log_dir to a new directory and checkpoint_path to the output path of the previous training results. If the output path is an OBS directory, set the path to a value starting with obs://.
+ If labels are changed for label data, perform operations in If Labels Are Changed before running mox.run.
+ mox.run(input_fn=input_fn,
+ model_fn=model_fn,
+ optimizer_fn=optimizer_fn,
+ run_mode=flags.run_mode,
+ inter_mode=mox.ModeKeys.EVAL if use_eval_data else None,
+ log_dir=log_dir,
+ batch_size=batch_size_per_device,
+ auto_batch=False,
+ max_number_of_steps=max_number_of_steps,
+ log_every_n_steps=flags.log_every_n_steps,
+ save_summary_steps=save_summary_steps,
+ save_model_secs=save_model_secs,
+ checkpoint_path=flags.checkpoint_url,
+ export_model=mox.ExportKeys.TF_SERVING)
+
+ If Labels Are ChangedIf the labels in a dataset have changed, execute the following statement. The statement must be executed before running mox.run.
+ In the statement, the logits variable indicates classification layer weights in different networks, and different parameters are configured. Set this parameter to the corresponding keyword.
+ mox.set_flag('checkpoint_exclude_patterns', 'logits')
+ If the built-in network of MoXing is used, the corresponding keyword needs to be obtained by calling the following API. In this example, the Resnet_v1_50 keyword is the value of logits.
+ import moxing.tensorflow as mox
+
+model_meta = mox.get_model_meta(mox.NetworkKeys.RESNET_V1_50)
+logits_pattern = model_meta.default_logits_pattern
+print(logits_pattern)
+ You can also obtain a list of networks supported by MoXing by calling the following API:
+ import moxing.tensorflow as mox
+print(help(mox.NetworkKeys))
+ The following information is displayed:
+ Help on class NetworkKeys in module
+moxing.tensorflow.nets.nets_factory:
+
+class NetworkKeys(builtins.object)
+ | Data descriptors defined here:
+ |
+ | __dict__
+ | dictionary for instance variables (if defined)
+ |
+ | __weakref__
+ | list of weak references to the object (if defined)
+ |
+ | ----------------------------------------------------------------------
+ | Data and other attributes defined here:
+ |
+ | ALEXNET_V2 = 'alexnet_v2'
+ |
+ | CIFARNET = 'cifarnet'
+ |
+ | INCEPTION_RESNET_V2 = 'inception_resnet_v2'
+ |
+ | INCEPTION_V1 = 'inception_v1'
+ |
+ | INCEPTION_V2 = 'inception_v2'
+ |
+ | INCEPTION_V3 = 'inception_v3'
+ |
+ | INCEPTION_V4 = 'inception_v4'
+ |
+ | LENET = 'lenet'
+ |
+ | MOBILENET_V1 = 'mobilenet_v1'
+ |
+ | MOBILENET_V1_025 = 'mobilenet_v1_025'
+ |
+ | MOBILENET_V1_050 = 'mobilenet_v1_050'
+ |
+ | MOBILENET_V1_075 = 'mobilenet_v1_075'
+ |
+ | MOBILENET_V2 = 'mobilenet_v2'
+ |
+ | MOBILENET_V2_035 = 'mobilenet_v2_035'
+ |
+ | MOBILENET_V2_140 = 'mobilenet_v2_140'
+ |
+ | NASNET_CIFAR = 'nasnet_cifar'
+ |
+ | NASNET_LARGE = 'nasnet_large'
+ |
+ | NASNET_MOBILE = 'nasnet_mobile'
+ |
+ | OVERFEAT = 'overfeat'
+ |
+ | PNASNET_LARGE = 'pnasnet_large'
+ |
+ | PNASNET_MOBILE = 'pnasnet_mobile'
+ |
+ | PVANET = 'pvanet'
+ |
+ | RESNET_V1_101 = 'resnet_v1_101'
+ |
+ | RESNET_V1_110 = 'resnet_v1_110'
+ |
+ | RESNET_V1_152 = 'resnet_v1_152'
+ |
+ | RESNET_V1_18 = 'resnet_v1_18'
+ |
+ | RESNET_V1_20 = 'resnet_v1_20'
+ |
+ | RESNET_V1_200 = 'resnet_v1_200'
+ |
+ | RESNET_V1_50 = 'resnet_v1_50'
+ |
+ | RESNET_V1_50_8K = 'resnet_v1_50_8k'
+ |
+ | RESNET_V1_50_MOX = 'resnet_v1_50_mox'
+ |
+ | RESNET_V1_50_OCT = 'resnet_v1_50_oct'
+ |
+ | RESNET_V2_101 = 'resnet_v2_101'
+ |
+ | RESNET_V2_152 = 'resnet_v2_152'
+ |
+ | RESNET_V2_200 = 'resnet_v2_200'
+ |
+ | RESNET_V2_50 = 'resnet_v2_50'
+ |
+ | RESNEXT_B_101 = 'resnext_b_101'
+ |
+ | RESNEXT_B_50 = 'resnext_b_50'
+ |
+ | RESNEXT_C_101 = 'resnext_c_101'
+ |
+ | RESNEXT_C_50 = 'resnext_c_50'
+ |
+ | VGG_16 = 'vgg_16'
+ |
+ | VGG_16_BN = 'vgg_16_bn'
+ |
+ | VGG_19 = 'vgg_19'
+ |
+ | VGG_19_BN = 'vgg_19_bn'
+ |
+ | VGG_A = 'vgg_a'
+ |
+ | VGG_A_BN = 'vgg_a_bn'
+ |
+ | XCEPTION_41 = 'xception_41'
+ |
+ | XCEPTION_65 = 'xception_65'
+ |
+ | XCEPTION_71 = 'xception_71'
+
+ How Do I View All Files Stored in OBS on ModelArts?
+To view all files stored in OBS when using notebook instances or training jobs, use either of the following methods:
+ - OBS console
Log in to OBS console using the current account, and search for the OBS buckets, folders, and files.
+ - You can use an API to check whether a given directory exists. In an existing notebook instance or after creating a new notebook instance, run the following command to check whether the directory exists:
import moxing as mox
+mox.file.list_directory('obs://bucket_name', recursive=True)
+If there are a large number of files, wait until the final file path is displayed.
+
+ How Do I Invoke a Shell Script in a Training Job to Execute the .sh File?
+ModelArts enables you to invoke a shell script, and you can use Python to invoke .sh. The procedure is as follows:
+ - Upload the .sh script to an OBS bucket. For example, upload the .sh script to /bucket-name/code/test.sh.
- Create the .py file on a local PC, for example, test.py. The background automatically downloads the code directory to the /home/work/user-job-dir/ directory of the container. Therefore, you can invoke the .sh file in the test.py boot file as follows:
import os
+os.system('bash /home/work/user-job-dir/code/test.sh')
+ - Upload test.py to OBS. Then the file storage path is /bucket-name/code/test.py.
- When creating a training job, set the code directory to /bucket-name/code/, and the boot file directory to /bucket-name/code/test.py.
+ After the training job is created, you can use Python to invoke the .sh file.
+ Where Are Datasets of ModelArts Stored in a Container?
+Datasets of ModelArts and data in specific data storage locations are stored in OBS.
+ Do Files in /cache Still Exist After a Notebook Instance is Stopped or Restarted? How Do I Avoid a Restart?
-/cache is a temporary directory and will not be saved. After an instance using OBS storage is stopped, data in the ~work directory will be deleted. After a notebook instance is restarted, all cached data except the data in the OBS bucket is lost, and your model or code is unavailable.
- To avoid a restart, do not train heavy-load jobs that consume large amounts of CPU, GPU, or memory resources in DevEnviron.
+ Temporary files are stored in the /cache directory and will not be saved after the notebook instance is stopped or restarted. Data stored in the /home/ma-user/work directory will be retained after the notebook instance is stopped or restarted.
+ To avoid a restart, do not train heavy-load jobs that consume large amounts of CPU, GPU, or memory resources in the development environment.
diff --git a/docs/modelarts/umn/modelarts_05_0082.html b/docs/modelarts/umn/modelarts_05_0082.html
new file mode 100644
index 00000000..7aa43f70
--- /dev/null
+++ b/docs/modelarts/umn/modelarts_05_0082.html
@@ -0,0 +1,30 @@
+
+
+ How Do I View GPU Usage on the Notebook?
+ If you select GPU when creating a notebook instance, perform the following operations to view GPU usage:
+ - Log in to the ModelArts management console, and choose .
- In the Operation column of the target notebook instance in the notebook list, click Open to go to the Jupyter page.
- On the Files tab page of the Jupyter page, click New and select Terminal. The Terminal page is displayed.
- Run the following command to view GPU usage:
nvidia-smi
+ - Check which processes in the current notebook instance use GPUs.
Method 1:
+python /modelarts/tools/gpu_processes.py
+The following figure shows the case that the current process is using GPUs.
+
+The following figure shows the case that the current process is not using GPUs.
+
+Method 2:
+Open /resource_info/gpu_usage.json and view the processes that are using GPUs.
+
+If no process is using GPUs, the file may be unavailable or empty.
+
+
+
+
+
+
\ No newline at end of file
diff --git a/docs/modelarts/umn/modelarts_05_0084.html b/docs/modelarts/umn/modelarts_05_0084.html
new file mode 100644
index 00000000..841b51e9
--- /dev/null
+++ b/docs/modelarts/umn/modelarts_05_0084.html
@@ -0,0 +1,11 @@
+
+
+ Does ModelArts Support the Caffe Engine?
+ The Python 2 environment of ModelArts supports Caffe, but the Python 3 environment does not support it.
+
+
+
diff --git a/docs/modelarts/umn/modelarts_05_0085.html b/docs/modelarts/umn/modelarts_05_0085.html
new file mode 100644
index 00000000..78cfe9be
--- /dev/null
+++ b/docs/modelarts/umn/modelarts_05_0085.html
@@ -0,0 +1,15 @@
+
+
+ How Do I Rename an OBS File?
+ OBS files cannot be renamed on the OBS console. To rename an OBS file, call a MoXing API in an existing or newly created notebook instance.
+ The following shows an example:
+ Rename obs_file.txt obs_file_2.txt. import moxing as mox
+mox.file.rename('obs://bucket_name/obs_file.txt', 'obs://bucket_name/obs_file_2.txt')
+
+
+
+
diff --git a/docs/modelarts/umn/modelarts_05_0086.html b/docs/modelarts/umn/modelarts_05_0086.html
new file mode 100644
index 00000000..c19c2c40
--- /dev/null
+++ b/docs/modelarts/umn/modelarts_05_0086.html
@@ -0,0 +1,11 @@
+
+
+ What Is the Common File Path for Training Jobs?
+ The path to the training environment and the code directory in the container are generally obtained using the environment variable ${MA_JOB_DIR}, which is /home/ma-user/modelarts/user-job-dir.
+
+
+
diff --git a/docs/modelarts/umn/modelarts_05_0088.html b/docs/modelarts/umn/modelarts_05_0088.html
index 7668a02a..882a9f5d 100644
--- a/docs/modelarts/umn/modelarts_05_0088.html
+++ b/docs/modelarts/umn/modelarts_05_0088.html
@@ -1,24 +1,24 @@
-
+
How Do I Install a Library That C++ Depends on?
- A third-party library may be used during job training. The following uses C++ as an example to describe how to install a third-party library.
- - Download source code to a local PC and upload it to OBS. For details about how to upload a file using OBS Browser, see .
- Use MoXing to copy the source code uploaded to OBS to a notebook instance in the development environment.
The following is a code example for copying data to a notebook instance in a development environment running on an EVS: import moxing as mox
+A third-party library may be used during job training. The following uses C++ as an example to describe how to install a third-party library.
+ - Download source code to a local PC and upload it to OBS. For details about how to upload a file using OBS Browser, see .
- Use MoXing to copy the source code uploaded to OBS to a notebook instance in the development environment.
The following is a code example for copying data to a notebook instance in a development environment running on an EVS: import moxing as mox
mox.file.make_dirs('/home/ma-user/work/data')
mox.file.copy_parallel('obs://bucket-name/data', '/home/ma-user/work/data')
- On the Files tab page of the Jupyter page, click New and select Terminal. Run the following command to go to the target path, and check whether the source code has been downloaded, that is, whether the data file exists.
cd /home/ma-user/work
+ - On the Files tab page of the Jupyter page, click New and select Terminal. Run the following command to go to the target path, and check whether the source code has been downloaded, that is, whether the data file exists.
cd /home/ma-user/work
ls
- - Compile code in Terminal based on service requirements.
- Use MoXing to copy the compilation results to OBS. The following is a code example.
import moxing as mox
+ - Compile code in Terminal based on service requirements.
- Use MoXing to copy the compilation results to OBS. The following is a code example.
import moxing as mox
mox.file.make_dirs('/home/ma-user/work/data')
mox.file.copy_parallel('/home/ma-user/work/data', 'obs://bucket-name/file)
- - During training, use MoXing to copy the compilation result from OBS to the container. The following is a code example.
import moxing as mox
+ - During training, use MoXing to copy the compilation result from OBS to the container. The following is a code example.
import moxing as mox
mox.file.make_dirs('/cache/data')
mox.file.copy_parallel('obs://bucket-name/data', '/cache/data')
How Do I Check Resource Usage of a Training Job?
+In the left navigation pane of the ModelArts management console, choose to go to the Training Jobs page. In the training job list, click a job name to view job details. You can view the following metrics on the Resource Usages tab page.
+ - CPU: CPU usage (cpuUsage) percentage (Percent)
- MEM: Physical memory usage (memUsage) percentage (Percent)
- GPU: GPU usage (gpuUtil) percentage (Percent)
- GPU_MEM: GPU memory usage (gpuMemUsage) percentage (Percent)
+ What Are Sizes of the /cache Directories for Different Resource Specifications in the Training Environment?
-When creating a training job, you can select CPU, GPU, or Ascend resources based on the size of the training job.
- ModelArts mounts the disk to the /cache directory. You can use this directory to store temporary files. The /cache directory shares resources with the code directory. The directory has different capacities for different resource specifications.
- - GPU resources
-
Table 1 Capacities of the cache directories for GPU resourcesGPU Specifications
+When creating a training job, you can select CPU, GPU, or Ascend resources based on the size of the training job.
+ ModelArts mounts a disk to /cache. You can use this directory to store temporary files. The /cache directory shares resources with the code directory. The directory has different capacities for different resource specifications.
+ - GPU resources
+
Table 1 Capacities of the cache directories for GPU resourcesGPU Specifications
|
-cache Directory Capacity
+ | cache Directory Capacity
|
|---|
-V100
+ | V100
|
-800G
+ | 800 GB
|
-8*V100
+ | 8*V100
|
-3T
+ | 3 TB
|
-P100
+ | P100
|
-800G
+ | 800 GB
|
- CPU resources
-
Table 2 Capacities of the cache directories for CPU resourcesCPU Specifications
+- CPU resources
+
Table 2 Capacities of the cache directories for CPU resources|
|
|---|
|
|---|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
in the filter criteria area and add filter criteria to quickly filter the data you want to view.
.)
