AzureML v1 (deepspeedai#185)

* remove * to include ds_config.json file in job

* add bookcorpus_data folder

* note for large files (docker build context)

* update docker image to upgrade deepspeed

Author: konabuta

.gitignore

@@ -12,9 +12,12 @@ dist/
 *.swp
 
 # AML workspace config file
-*config.json
+config.json
 
 .coverage_*
 *~
 slurm*
 logs
+
+# Data folder
+bookcorpus_data/

examples_deepspeed/azureml/Dockerfile.dockerfile

@@ -1,4 +1,4 @@
-FROM mcr.microsoft.com/azureml/curated/acpt-pytorch-1.11-py38-cuda11.5-gpu
+FROM mcr.microsoft.com/azureml/curated/acpt-pytorch-1.11-cuda11.3:12
 USER root:root
 
 RUN pip install pybind11

examples_deepspeed/azureml/README.md

@@ -10,5 +10,7 @@ Setup an AML workspace. Refer to: [set-up doc](https://github.com/Azure/azureml-
 Create AML Dataset. To run remote AML job, you need to provide AML FileDataset. 
 Refer to [prepare_dataset script](prepare_dataset.py) to upload .bin and .idx files to blob store and on how to create FileDataset.
 
+> Note: The folder `bookcorpus_data` used by [prepare_dataset script](prepare_dataset.py) should not be under `azureml` directories. It is because Azure ML does not allow to include large files (limit: 100 files or 1048576 bytes) for Docker build context.
+
 # Training
 Run Megatron-DeepSpeed on Azure ML. Refer to [aml_submit script](aml_submit.py).