Keep --max-model-len 2048 so vLLM fits on a 24 GB GPU

README.md

@@ -48,7 +48,7 @@ bidirectional InfoNCE loss.
 pip install contrastive-lm
 
 # 1. encoder (Qwen3-8B embeddings)
-vllm serve Qwen/Qwen3-8B --served-model-name qwen3-8b --runner pooling --port 8090 &
+vllm serve Qwen/Qwen3-8B --served-model-name qwen3-8b --runner pooling --max-model-len 2048 --port 8090 &
 
 # 2. API + playground at http://localhost:8700/ (fetches CLM_v0.1-8B.pt into ~/.cache/clm/)
 clm-serve