Update README.md

README.md

@@ -104,7 +104,7 @@ print("generate_text:", generate_text)
 
 ### Using vLLM
 
-vLLM is currently being adapted, priority can be given to using our forked repository [vllm](https://github.com/CSWYF3634076/vllm/tree/ernie). We are working with the community to fully support ERNIE4.5 models, stay tuned.
+[vllm](https://github.com/vllm-project/vllm/tree/main) github library. Python-only [build](https://docs.vllm.ai/en/latest/getting_started/installation/gpu.html#set-up-using-python-only-build-without-compilation).
 
 ```bash
 # 80G * 16 GPU
@@ -112,7 +112,7 @@ vllm serve baidu/ERNIE-4.5-300B-A47B-PT --trust-remote-code
 ```
 
 ```bash
-# FP8 online quantification 80G * 8 GPU
+# FP8 online quantification 80G * 16 GPU
 vllm serve baidu/ERNIE-4.5-300B-A47B-PT --trust-remote-code --quantization fp8
 ```