Merge branch 'master' into master

pytorch · May 10, 2023 · 12f7d48 · 12f7d48
2 parents 0d8cebf + fbcde1d
commit 12f7d48
Show file tree

Hide file tree

Showing 3 changed files with 2 additions and 53 deletions.
diff --git a/docs/large_model_inference.md b/docs/large_model_inference.md
@@ -181,7 +181,7 @@ torch-model-archiver --model-name bloom --version 1.0 --handler deepspeed_handle
 * Pre-install the model parallel library such as Deepspeed on the container or host.
 * Pre-download the model checkpoints. For example, if using HuggingFace pretrained model can be pre-downloaded via [Download_model.py](https://github.com/pytorch/serve/blob/75f66dc557b3b67a3ab56536a37d7aa21582cc04/examples/large_models/deepspeed/opt/Readme.md?plain=1#L7)
   * Set environment variable [HUGGINGFACE_HUB_CACHE](https://huggingface.co/docs/huggingface_hub/guides/manage-cache#understand-caching) and [TRANSFORMERS_CACHE](https://huggingface.co/transformers/v4.0.1/installation.html#caching-models)
-  * Download model to the HuggingFace cache dir via tool [Download_model.py](https://github.com/pytorch/serve/blob/75f66dc557b3b67a3ab56536a37d7aa21582cc04/examples/large_models/deepspeed/opt/Readme.md?plain=1#L7)
+  * Download model to the HuggingFace cache dir via tool [Download_model.py](https://github.com/pytorch/serve/blob/4fe5273cd6f98fb5abc570f802b402ac32ecd105/examples/large_models/Huggingface_pippy/Readme.md?plain=1#L20)
 
 #### Tune "[responseTimeout](https://github.com/pytorch/serve/blob/5ee02e4f050c9b349025d87405b246e970ee710b/docs/configuration.md?plain=1#L216)" (see [model config YAML file](https://github.com/pytorch/serve/blob/5ee02e4f050c9b349025d87405b246e970ee710b/model-archiver/README.md?plain=1#L164)) if high model loading or inference latency causes response timeout.
 

diff --git a/examples/large_models/deepspeed/opt/Readme.md b/examples/large_models/deepspeed/opt/Readme.md
@@ -7,7 +7,7 @@ This document briefs on serving large HG models on multiple GPUs using deepspeed
 ### Step 1: Download model
 
 ```bash
-python ../../utils/Download_models.py --model_path model --model_name facebook/opt-350m --revision main
+python ../../utils/Download_model.py --model_path model --model_name facebook/opt-350m --revision main
 ```
 
 The script prints the path where the model is downloaded as below.

diff --git a/examples/large_models/utils/Download_models.py b/examples/large_models/utils/Download_models.py