Instructions to use selimaktas/MiniMax-M2.75-460B-A20B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use selimaktas/MiniMax-M2.75-460B-A20B with Transformers:

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

pipe = pipeline("text-generation", model="selimaktas/MiniMax-M2.75-460B-A20B", trust_remote_code=True)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("selimaktas/MiniMax-M2.75-460B-A20B", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("selimaktas/MiniMax-M2.75-460B-A20B", trust_remote_code=True)

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use selimaktas/MiniMax-M2.75-460B-A20B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "selimaktas/MiniMax-M2.75-460B-A20B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "selimaktas/MiniMax-M2.75-460B-A20B",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/selimaktas/MiniMax-M2.75-460B-A20B

SGLang

How to use selimaktas/MiniMax-M2.75-460B-A20B 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 "selimaktas/MiniMax-M2.75-460B-A20B" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "selimaktas/MiniMax-M2.75-460B-A20B",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

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 "selimaktas/MiniMax-M2.75-460B-A20B" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "selimaktas/MiniMax-M2.75-460B-A20B",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use selimaktas/MiniMax-M2.75-460B-A20B with Docker Model Runner:
```
docker model run hf.co/selimaktas/MiniMax-M2.75-460B-A20B
```

How to deploy?

by LetsJumP - opened Apr 22

Discussion

LetsJumP

Apr 22

Seems great!
Any deployment guidance?

selimaktas

Owner Apr 22

works with just like minimax itself vLLM, you might need to pass --tokenizer MiniMaxAI/MiniMax-M2.7 for it to work, but might also work without, i haven't tested that!
For 8xRTX PRO 6000 I use:

SAFETENSORS_FAST_GPU=1 \
vllm serve selimaktas/MiniMax-M2.75-460B-A20B \
    --tokenizer MiniMaxAI/MiniMax-M2.7 \
    --tensor-parallel-size 4 \
    --data-parallel-size 2 \
    --enable-expert-parallel \
    --gpu-memory-utilization 0.92 \
    --enable-auto-tool-choice \
    --tool-call-parser minimax_m2 \
    --reasoning-parser minimax_m2 \
    --trust-remote-code

lkjrkljlwladw

about 1 month ago

any possibility for nvfp4?

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