Instructions to use QuantTrio/GLM-5.2-Int4-Int8Mix with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use QuantTrio/GLM-5.2-Int4-Int8Mix with Transformers:

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

pipe = pipeline("text-generation", model="QuantTrio/GLM-5.2-Int4-Int8Mix")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForMultimodalLM

tokenizer = AutoTokenizer.from_pretrained("QuantTrio/GLM-5.2-Int4-Int8Mix")
model = AutoModelForMultimodalLM.from_pretrained("QuantTrio/GLM-5.2-Int4-Int8Mix")
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/GLM-5.2-Int4-Int8Mix with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "QuantTrio/GLM-5.2-Int4-Int8Mix"
# 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/GLM-5.2-Int4-Int8Mix",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/QuantTrio/GLM-5.2-Int4-Int8Mix

SGLang

How to use QuantTrio/GLM-5.2-Int4-Int8Mix 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/GLM-5.2-Int4-Int8Mix" \
    --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/GLM-5.2-Int4-Int8Mix",
		"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/GLM-5.2-Int4-Int8Mix" \
        --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/GLM-5.2-Int4-Int8Mix",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use QuantTrio/GLM-5.2-Int4-Int8Mix with Docker Model Runner:
```
docker model run hf.co/QuantTrio/GLM-5.2-Int4-Int8Mix
```

AWQ 4bit

by MatthieuZ - opened about 7 hours ago

Discussion

MatthieuZ

about 7 hours ago

Any plan to make your usual AWQ 4bit version of this model?

tclf90

QuantTrio org about 6 hours ago

•

edited about 4 hours ago

Thanks for asking!

For now I don’t plan to make a separate “classic AWQ 4-bit” release.

One reason is that my recent AWQ-style releases have gradually moved away from data calibration. If we put the calibration strategy aside, the remaining part is mostly the packing/runtime format, which is essentially WnA16: low-bit weights with 16-bit activations. So format-wise, I’m now moving more toward a generic WnA16 / compressed-tensors representation rather than the older AWQ format.

Also, vLLM’s support for compressed-tensors has become much better, and it gives more fine-grained control over which parts to be kept in int4 / int8 / other formats. I’ve also seen people built FP4-Int8Mix variants based on this repo, so this format is quite convenient for experimentation.

If you specifically need the old AWQ format for compatibility with a certain backend, you can probably ask Codex / Claude Code to help repack this repo directly into an AWQ format. This is a pretty straightforward repacking task, and it should work out right without an issue.

MatthieuZ

about 4 hours ago

Ok thanks for the hint, we'll try to repack.

MatthieuZ changed discussion status to closed about 4 hours ago

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