Update README.md

README.md

@@ -7,6 +7,16 @@ base_model:
 - Qwen/Qwen3-30B-A3B-Base
 ---
 
+# Sharded weights checkpoints
+
+This is derived directly from [`save_sharded_state.py`](https://github.com/vllm-project/vllm/blob/main/examples/offline_inference/save_sharded_state.py) to be used with vLLM with `-tp=2`:
+
+```bash
+vllm serve aarnphm/qwen3-30b-a3b-sharded-tp2 \
+      -tp=2 \
+      --load-format sharded_state
+```
+
 # Qwen3-30B-A3B
 <a href="https://chat.qwen.ai/" target="_blank" style="margin: 2px;">
     <img alt="Chat" src="https://img.shields.io/badge/%F0%9F%92%9C%EF%B8%8F%20Qwen%20Chat%20-536af5" style="display: inline-block; vertical-align: middle;"/>