Instructions to use bullerwins/MiniMax-M3-4bit-W4A16-v0 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use bullerwins/MiniMax-M3-4bit-W4A16-v0 with Transformers:

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

pipe = pipeline("image-text-to-text", model="bullerwins/MiniMax-M3-4bit-W4A16-v0", trust_remote_code=True)
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
pipe(text=messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoProcessor, AutoModelForMultimodalLM

processor = AutoProcessor.from_pretrained("bullerwins/MiniMax-M3-4bit-W4A16-v0", trust_remote_code=True)
model = AutoModelForMultimodalLM.from_pretrained("bullerwins/MiniMax-M3-4bit-W4A16-v0", trust_remote_code=True, device_map="auto")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
inputs = processor.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(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use bullerwins/MiniMax-M3-4bit-W4A16-v0 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "bullerwins/MiniMax-M3-4bit-W4A16-v0"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "bullerwins/MiniMax-M3-4bit-W4A16-v0",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker

docker model run hf.co/bullerwins/MiniMax-M3-4bit-W4A16-v0

SGLang

How to use bullerwins/MiniMax-M3-4bit-W4A16-v0 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 "bullerwins/MiniMax-M3-4bit-W4A16-v0" \
    --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": "bullerwins/MiniMax-M3-4bit-W4A16-v0",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

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 "bullerwins/MiniMax-M3-4bit-W4A16-v0" \
        --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": "bullerwins/MiniMax-M3-4bit-W4A16-v0",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Docker Model Runner
How to use bullerwins/MiniMax-M3-4bit-W4A16-v0 with Docker Model Runner:
```
docker model run hf.co/bullerwins/MiniMax-M3-4bit-W4A16-v0
```

bullerwins commited on Jun 13

Commit

6c09f6a

verified ·

1 Parent(s): ef5f89e

Update README.md

Browse files

Files changed (1) hide show

README.md +12 -6

README.md CHANGED Viewed

@@ -5,13 +5,19 @@ license_name: minimax-community
 license_link: LICENSE
 library_name: transformers
 tags:
-  - multimodal
-  - moe
-  - agent
-  - coding
-  - video
 ---
 <div align="center">
   <img width="60%" src="figures/logo.svg" alt="MiniMax">
 </div>
@@ -84,4 +90,4 @@ We recommend the following parameters for best performance: `temperature=1.0`, `
 ## Contact Us
-Contact us at [model@minimax.io](mailto:model@minimax.io).

 license_link: LICENSE
 library_name: transformers
 tags:
+- multimodal
+- moe
+- agent
+- coding
+- video
+base_model:
+- MiniMaxAI/MiniMax-M3
 ---
+Experimental int4 w4a16, I have not been able to test it as the vLLM M3's support PR does not support pipeline paralelism and I don't have the hardware to test tensor paralelism, so here may be dragons, but people like to tinker.
+You will need this PR from vllm to make it work https://github.com/vllm-project/vllm/pull/45381
+This is using RTN quantization, not full calibrated Auto-round.
 <div align="center">
   <img width="60%" src="figures/logo.svg" alt="MiniMax">
 </div>
 ## Contact Us
+Contact us at [model@minimax.io](mailto:model@minimax.io).