Instructions to use QuantTrio/Qwen3-Coder-30B-A3B-Instruct-AWQ with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use QuantTrio/Qwen3-Coder-30B-A3B-Instruct-AWQ with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="QuantTrio/Qwen3-Coder-30B-A3B-Instruct-AWQ")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("QuantTrio/Qwen3-Coder-30B-A3B-Instruct-AWQ")
model = AutoModelForCausalLM.from_pretrained("QuantTrio/Qwen3-Coder-30B-A3B-Instruct-AWQ", device_map="auto")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use QuantTrio/Qwen3-Coder-30B-A3B-Instruct-AWQ with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "QuantTrio/Qwen3-Coder-30B-A3B-Instruct-AWQ"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "QuantTrio/Qwen3-Coder-30B-A3B-Instruct-AWQ",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/QuantTrio/Qwen3-Coder-30B-A3B-Instruct-AWQ

SGLang

How to use QuantTrio/Qwen3-Coder-30B-A3B-Instruct-AWQ with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "QuantTrio/Qwen3-Coder-30B-A3B-Instruct-AWQ" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "QuantTrio/Qwen3-Coder-30B-A3B-Instruct-AWQ",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "QuantTrio/Qwen3-Coder-30B-A3B-Instruct-AWQ" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "QuantTrio/Qwen3-Coder-30B-A3B-Instruct-AWQ",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use QuantTrio/Qwen3-Coder-30B-A3B-Instruct-AWQ with Docker Model Runner:
```
docker model run hf.co/QuantTrio/Qwen3-Coder-30B-A3B-Instruct-AWQ
```

INT4 KV cache in LMDeploy breaks tool calling

by zambalee-ai - opened Jul 8

Discussion

zambalee-ai

Jul 8

Issue: INT4 KV cache (`--quant-policy 4`) breaks `qwen3coder` tool call parser for Qwen3-Coder-30B

When deploying this model with LMDeploy 0.14.0 using INT4 KV cache quantization, the qwen3coder tool call parser fails to extract structured tool calls from model output. The same model with INT8 KV cache works perfectly.

Details

LMDeploy: 0.14.0
Hardware: 4xV100-SXM2-16GB (tp=4)
INT8 KV: tool_calls returned correctly, MTCSR chain_success 5/6 (83.3%)
INT4 KV: tool_calls returns empty [], MTCSR chain_success 0/6

Root cause

INT4 KV cache quantization causes the model to stop emitting the expected XML tool call format. The parser cannot find the expected pattern and returns empty tool_calls.

Workaround

Use INT8 KV cache (--quant-policy 8) instead. This works correctly but uses more GPU memory.

LMDeploy issue: https://github.com/InternLM/lmdeploy/issues/4743
The issue may be related to how INT4 KV affects the model's attention computation, causing it to output plain text instead of structured XML tool calls.

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment

QuantTrio
/

Qwen3-Coder-30B-A3B-Instruct-AWQ

INT4 KV cache in LMDeploy breaks tool calling

Issue: INT4 KV cache (`--quant-policy 4`) breaks `qwen3coder` tool call parser for Qwen3-Coder-30B

Details

Root cause

Workaround

Related

INT4 KV cache in LMDeploy breaks tool calling

Issue: INT4 KV cache (--quant-policy 4) breaks qwen3coder tool call parser for Qwen3-Coder-30B

Details

Root cause

Workaround

Related

Issue: INT4 KV cache (`--quant-policy 4`) breaks `qwen3coder` tool call parser for Qwen3-Coder-30B