x

Dockerfile

@@ -4,9 +4,6 @@ FROM docker.1ms.run/kyuz0/vllm-therock-gfx1201:latest
 ENV USE_DEFAULT_MODEL=true
 ENV LOCAL_MODEL_DIR=/opt/model
 
-# Ensure latest Transformers model definitions are available
-RUN /opt/venv/bin/pip install --no-cache-dir --upgrade "git+https://github.com/huggingface/transformers.git"
-
 # Create necessary directories
 RUN mkdir -p /opt/script /opt/model /config
 

config.json

@@ -1,5 +1,5 @@
 {
-  "default_model": "Qwen3.5-35B-A3B-GPTQ-Int4",
+  "default_model": "Qwen3-Coder-30B-A3B-Instruct-GPTQ-4bit",
   "models": {
     "Meta-Llama-3.1-8B-Instruct": {
       "ctx": "65536",
@@ -106,7 +106,6 @@
       "max_num_seqs": "64",
       "max_tokens": "32768",
       "gpu_util": "0.98",
-      "model_impl": "transformers",
       "tool_call_parser": "qwen3_xml",
       "enable_auto_tool_choice": true,
       "served_model_name": "Qwen3.5-27B-FP8",
@@ -119,7 +118,6 @@
       "max_num_seqs": "64",
       "max_tokens": "32768",
       "gpu_util": "0.98",
-      "model_impl": "transformers",
       "tool_call_parser": "qwen3_xml",
       "enable_auto_tool_choice": true,
       "served_model_name": "Qwen3.5-35B-A3B-GPTQ-Int4",

scripts/start_vllm.py

@@ -144,11 +144,6 @@ def launch_model(model_id, config, model_path, gpu_count):
         "--gpu-memory-utilization", str(gpu_util),
         "--dtype", "auto"
     ]
-
-    model_impl = config.get("model_impl") or os.getenv("VLLM_MODEL_IMPL")
-    if model_impl:
-        cmd.extend(["--model-impl", str(model_impl)])
-        log(f"Added model implementation: {model_impl}")
     
     if config.get("trust_remote"):
         cmd.append("--trust-remote-code")