Upload qwen2.5-7b.env with huggingface_hub

qwen2.5-7b.env

@@ -0,0 +1,12 @@
+# Qwen 2.5 7B Instruct — original baseline recipe
+MODEL=Qwen/Qwen2.5-7B-Instruct
+SERVED_NAME=Qwen/Qwen2.5-7B-Instruct
+IMAGE=nvcr.io/nvidia/vllm:26.04-py3
+GPU_MEM_UTIL=0.85
+
+EXTRA_DOCKER_ARGS=()
+
+EXTRA_VLLM_ARGS=(
+  --dtype bfloat16
+  --max-model-len 8192
+)