Update vLLM install instructions in README

README.md

@@ -251,10 +251,23 @@ More prompt recipes are available in the GitHub repository: <https://github.com/
 
 ### Serve with vLLM and SGLang
 
-MOSS-Transcribe-Diarize 0.9B supports vLLM serving through the OpenAI-compatible transcription API:
+MOSS-Transcribe-Diarize supports vLLM serving through the OpenAI-compatible transcription API. Use a pinned vLLM nightly build that includes the MOSS-Transcribe-Diarize model registration. Choose one of the following commands: for CUDA 12 environments, use `cu129`; for CUDA 13 environments, use `cu130`.
+
+```bash
+uv pip install -U vllm \
+  --torch-backend=auto \
+  --extra-index-url https://wheels.vllm.ai/68b4a1d582818e67adc903bf1b8fc5a5447da2fa/cu129
+```
+
+or:
+
+```bash
+uv pip install -U vllm \
+  --torch-backend=auto \
+  --extra-index-url https://wheels.vllm.ai/68b4a1d582818e67adc903bf1b8fc5a5447da2fa/cu130
+```
 
 ```bash
-pip install vllm
 vllm serve OpenMOSS-Team/MOSS-Transcribe-Diarize --trust-remote-code
 ```