diff --git a/examples/rinna-gpt2-train/cli/README.md b/examples/rinna-gpt2-train/cli/README.md
index cae5c41..ecd7a60 100644
--- a/examples/rinna-gpt2-train/cli/README.md
+++ b/examples/rinna-gpt2-train/cli/README.md
@@ -10,12 +10,19 @@ login to Azure
 az login
 ```
 
+check Azure subscription
+```bash
+az account show -o table # check if you are logged in to a right Azure subscription
+az account set --subscription <subscription_id> # change Azure subscription if needed
+```
+
 ## Installation Azure ML CLI 2.0 (Preview)
 
 Install Azure ML CLI 2.0 (Preview)
 
 ```bash
 ./setup.sh
+# Inside the shell, run the following command:
 # az extension add -n ml -y   <-- If you use devcontainer, it already installed. Seee Dockerfile under .devcontiner folder.
 # az configure --defaults group="azureml-mlops" workspace="azureml-mlops"  <-- change resource group name and workspace name to your environment.
 ```
@@ -34,6 +41,11 @@ Create Azure ML Environments based on [Dockerfile](./Dockerfile)
 Fine Tune GPT-2 Model
 
 ```bash
-./train.sh  # Fine Tune using HaggingFace Transformers Trainer API
+# single node (use transformers Trainer API directly)
+./train.sh  # Fine Tune rinna GPT-2 Model
 ```
-
+or 
+```bash
+# distributed training (use modified run_clm.py)
+./train_distributed.sh  # Fine Tune rinna GPT-2 Model
+```
\ No newline at end of file
diff --git a/examples/rinna-gpt2-train/cli/train_distributed.sh b/examples/rinna-gpt2-train/cli/train_distributed.sh
new file mode 100644
index 0000000..88a7a99
--- /dev/null
+++ b/examples/rinna-gpt2-train/cli/train_distributed.sh
@@ -0,0 +1,2 @@
+#!/bin/sh
+az ml job create --file rinna_trainer_distributed_job.yml