[vllm] Update README: fix issue #12 and others

vllm/README.md

@@ -1,36 +1,29 @@
-#  Image list
+# Image list
 
+- kuae1.3.0 : `registry.mthreads.com/mcconline/mtt-vllm-public:v0.2.0-kuae1.3.0`
 - kuae1.2.0 : `registry.mthreads.com/mcctest/musa-pytorch-transformer-vllm:v0.1.4-kuae1.2`
-- kuae1.3.0 : `registry.mthreads.com/mcctest/musa-pytorch-transformer-vllm:v0.1.4-kuae1.3`
 
 > "kuae1.x.0" is the driver version, you can check your driver version with the command `clinfo | grep 'Driver Version'`
 >
 > We have only verified the above images on the S4000 machine
 
-
-
 # Check vLLM Image
 
-Assume that your host environment has been configured successfully via [setup_musa](../setup_musa), including the driver, mt-container-toolkit: 
+Assume that your host environment has been configured successfully via [setup_musa](../setup_musa), including the driver, mt-container-toolkit:
 
 ```bash
 # (host env)
 # Usage: bash ./check_vllm_image/vllm_check.sh <driver_version>
 bash ./check_vllm_image/vllm_check.sh 1.2
 ```
 
-
-
 # Convert Weight
 
 ```bash
 bash ./convert_weight.sh -ckpt <model_path> -tp <tensor-parallel-size>
 ```
 
 > <convert_model_dir>: generated from the same level of <model_path>
-> 
-
-
 
 # Inference Demo
 
@@ -40,15 +33,10 @@ python generate_chat.py -ckpt <model_path>
 
 more demos(base on kuae1.2 driver version): [models](./models)
 
-
-
 # Benchmarks
 
 [Benchmarks](./benchmarks)
 
-
-
-#  More
+# More
 
 [MT-Transformer-vLLM User Guide](https://docs.mthreads.com/mtt/mtt-doc-online/)
-