sagemaker fixes and improvements (#708)

* adding aws sagemaker examples to examples readme

* refactoring and correcting documentation

docs/source/usage_guides/sagemaker.mdx

@@ -129,7 +129,26 @@ You can find your model data at: s3://your-bucket/accelerate-sagemaker-1-2021-04
 
 ### Distributed Training: Data Parallelism
 
-*currently in development, will be supported soon.*
+Set up the accelerate config by running `accelerate config` and answer the SageMaker questions and set it up.
+To use SageMaker DDP, select it when asked 
+`What is the distributed mode? ([0] No distributed training, [1] data parallelism):`.
+Example config below:
+```yaml
+base_job_name: accelerate-sagemaker-1
+compute_environment: AMAZON_SAGEMAKER
+distributed_type: DATA_PARALLEL
+ec2_instance_type: ml.p3.16xlarge
+iam_role_name: xxxxx
+image_uri: null
+mixed_precision: fp16
+num_machines: 1
+profile: xxxxx
+py_version: py38
+pytorch_version: 1.10.2
+region: us-east-1
+transformers_version: 4.17.0
+use_cpu: false
+```
 
 ### Distributed Training: Model Parallelism
 

examples/README.md

@@ -187,6 +187,9 @@ To run it in each of these various modes, use the following commands:
 ### Simple vision example (GANs)
 
 - [huggan project](https://github.com/huggingface/community-events/tree/main/huggan)
+
+### Using AWS SageMaker integration
+- [Examples showcasing AWS SageMaker integration of 🤗 Accelerate.](https://github.com/pacman100/accelerate-aws-sagemaker)
 
 ## Finer Examples
 

src/accelerate/commands/config/sagemaker.py

@@ -157,7 +157,7 @@ def get_sagemaker_input():
         )
 
     distributed_type = _ask_field(
-        "Which type of machine are you using? ([0] No distributed training, [1] data parallelism): ",
+        "What is the distributed mode? ([0] No distributed training, [1] data parallelism): ",
         _convert_sagemaker_distributed_mode,
         error_message="Please enter 0 or 1",
     )