Instructions to use maywell/GLM-4.5-Air-GLM-4.6-Distill with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use maywell/GLM-4.5-Air-GLM-4.6-Distill with Transformers:

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

pipe = pipeline("text-generation", model="maywell/GLM-4.5-Air-GLM-4.6-Distill")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("maywell/GLM-4.5-Air-GLM-4.6-Distill")
model = AutoModelForCausalLM.from_pretrained("maywell/GLM-4.5-Air-GLM-4.6-Distill", 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=256)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use maywell/GLM-4.5-Air-GLM-4.6-Distill with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "maywell/GLM-4.5-Air-GLM-4.6-Distill"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "maywell/GLM-4.5-Air-GLM-4.6-Distill",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/maywell/GLM-4.5-Air-GLM-4.6-Distill

SGLang

How to use maywell/GLM-4.5-Air-GLM-4.6-Distill 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 "maywell/GLM-4.5-Air-GLM-4.6-Distill" \
    --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": "maywell/GLM-4.5-Air-GLM-4.6-Distill",
		"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 "maywell/GLM-4.5-Air-GLM-4.6-Distill" \
        --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": "maywell/GLM-4.5-Air-GLM-4.6-Distill",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use maywell/GLM-4.5-Air-GLM-4.6-Distill with Docker Model Runner:
```
docker model run hf.co/maywell/GLM-4.5-Air-GLM-4.6-Distill
```

"Real SVD distill result" > with trash outputs

by mindkrypted - opened Oct 11, 2025

Discussion

mindkrypted

Oct 11, 2025

fantastic trolling attempt follow-up to BaseBased vibed-"SVD distill".

For those that are curious, it's a non-functional model, don't waste your bandwidth and compute.

maywell

Owner Oct 11, 2025

i told ya bro

ConicCat

Oct 11, 2025

Not even a troll, this is what BasedBase's models would have been like if the 'distill' merge method was actually applied properly.

ConicCat

Oct 11, 2025

Legitimately impressive that you got something like this to work at all though @maywell even if the results aren't that great.

mindkrypted

Oct 11, 2025

I consider this a trolling attempt as the story behind it isn't disclosed, yes, there's a screenshot of garbled text, but that's it.
More could be done to inform the community what is the model, what to expect and why not sharing insights behind what the code attempted to do~
In current state, this is pollution.

maywell

Owner Oct 11, 2025

I consider this a trolling attempt as the story behind it isn't disclosed, yes, there's a screenshot of garbled text, but that's it.
More could be done to inform the community what is the model, what to expect and why not sharing insights behind what the code attempted to do~
In current state, this is pollution.

I completely understand why you'd say that. I mentioned the context on Reddit and included the Hugging Face model address in the post, so I didn't provide a full explanation on the model card.
I've just added a link to the original Reddit post to the model card now.

gghfez

Oct 12, 2025

I found it useful, saved me wasting time/attention on the "SVD distill" rabbit hole.

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