Update README.md

README.md

@@ -82,6 +82,7 @@ The `embedl-models` package is required, it provides the optimized FlashHead imp
 ---
 
 ## Usage Examples
+**Note (vLLM context length):** `max_model_len=131072` may fail on GPUs without enough free VRAM for the KV cache. If you see a KV cache memory error, lower `max_model_len` (or increase `gpu_memory_utilization`).
 
 ### vLLM Inference
 
@@ -93,7 +94,7 @@ model_id = "embedl/Llama-3.2-3B-Instruct-FlashHead-W4A16"
 
 if __name__ == "__main__":
     sampling = SamplingParams(max_tokens=128, temperature=0.0)
-    llm = LLM(model=model_id, trust_remote_code=True)
+    llm = LLM(model=model_id, trust_remote_code=True, max_model_len=131072)
     
     prompt = "Write a haiku about coffee."
     output = llm.generate([prompt], sampling)
@@ -116,7 +117,8 @@ model_id = "embedl/Llama-3.2-3B-Instruct-FlashHead-W4A16"
 if __name__ == "__main__":
     asyncio.run(
         run_repl(
-            model=model_id
+            model=model_id,
+            max_model_len=131072
         )
     )
 ```