[examples] polish AutoParallel readme (#3270)

examples/tutorial/auto_parallel/README.md

@@ -45,6 +45,7 @@ colossalai run --nproc_per_node 4 auto_parallel_with_resnet.py
 You should expect to the log like this. This log shows the edge cost on the computation graph as well as the sharding strategy for an operation. For example, `layer1_0_conv1 S01R = S01R X RR` means that the first dimension (batch) of the input and output is sharded while the weight is not sharded (S means sharded, R means replicated), simply equivalent to data parallel training.
 ![](https://raw.githubusercontent.com/hpcaitech/public_assets/main/examples/tutorial/auto-parallel%20demo.png)
 
+**Note: This experimental feature has been tested on torch 1.12.1 and transformer 4.22.2. If you are using other versions, you may need to modify the code to make it work.**
 
 ### Auto-Checkpoint Tutorial
 

examples/tutorial/auto_parallel/requirements.txt

-torch
+torch==1.12.1
 colossalai
 titans
 pulp
 datasets
 matplotlib
-transformers
+transformers==4.22.1