Instructions to use QuantTrio/DeepSeek-V3.1-AWQ-Lite with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use QuantTrio/DeepSeek-V3.1-AWQ-Lite with Transformers:

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

pipe = pipeline("text-generation", model="QuantTrio/DeepSeek-V3.1-AWQ-Lite", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("QuantTrio/DeepSeek-V3.1-AWQ-Lite", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("QuantTrio/DeepSeek-V3.1-AWQ-Lite", trust_remote_code=True, 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/DeepSeek-V3.1-AWQ-Lite with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "QuantTrio/DeepSeek-V3.1-AWQ-Lite"
# 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/DeepSeek-V3.1-AWQ-Lite",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/QuantTrio/DeepSeek-V3.1-AWQ-Lite

SGLang

How to use QuantTrio/DeepSeek-V3.1-AWQ-Lite 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/DeepSeek-V3.1-AWQ-Lite" \
    --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/DeepSeek-V3.1-AWQ-Lite",
		"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/DeepSeek-V3.1-AWQ-Lite" \
        --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/DeepSeek-V3.1-AWQ-Lite",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use QuantTrio/DeepSeek-V3.1-AWQ-Lite with Docker Model Runner:
```
docker model run hf.co/QuantTrio/DeepSeek-V3.1-AWQ-Lite
```

random atrifacts on larger outputs

by willfalco - opened Dec 12, 2025

Discussion

willfalco

Dec 12, 2025

Random 极 symbols and sometimes others is this a result of "Light" or vllm inference?
...
For seconds >= 60, returns (minutes_digit, seconds_digit)
where minutes极digit is the base-60 digit for minutes
"""
total_seconds = int(seconds)
# For < 60 seconds, just return single digit
if total_seconds < 60:
return None, BASE60_DIGITS[total_seconds]
# For >= 60 seconds, calculate minutes and seconds in base-60
minutes = total_seconds // 60
seconds_remainder = total_seconds % 60
# Convert both to base-60 digits
if minutes < 极:
minutes_digit = BASE60_DIGITS[minutes]
else:
...

willfalco

Dec 12, 2025

though it gives better accuracy preservation then int4 (https://huggingface.co/QuantTrio/DeepSeek-V3.1-AWQ-Lite/discussions/3)
QuantTrio/DeepSeek-V3.1-AWQ-Lite Nonthinking 1 timeout (0/1) (vs 83.7% full posted)
business 86/789 wrong (89.1% accuracy)
law 394/1101 wrong (64.2% accuracy)
psychology 131/798 wrong (83.6% accuracy)
biology 71/717 wrong (90.1% accuracy)
chemistry 137/1132 wrong (87.9% accuracy)
history 96/381 wrong (74.8% accuracy)
other 180/924 wrong (80.5% accuracy)
health 170/818 wrong (79.2% accuracy)
economics 109/844 wrong (87.1% accuracy)
math 101/1351 wrong (92.5% accuracy)
physics 156/1299 wrong (88.0% accuracy)
computer science 53/410 wrong (87.1% accuracy)
philosophy 102/499 wrong (79.6% accuracy)
engineering 209/969 wrong (78.4% accuracy)

ALL CATEGORIES 1995/12032 wrong (83.4% accuracy)

willfalco

Dec 12, 2025

looks like some flashinfer sampler issue VLLM_USE_FLASHINFER_SAMPLER=1
but is ran with VLLM_ATTENTION_BACKEND=TRITON_MLA

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