Instructions to use saricles/MiniMax-M2.7-NVFP4-GB10 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use saricles/MiniMax-M2.7-NVFP4-GB10 with Transformers:

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

pipe = pipeline("text-generation", model="saricles/MiniMax-M2.7-NVFP4-GB10", 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("saricles/MiniMax-M2.7-NVFP4-GB10", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("saricles/MiniMax-M2.7-NVFP4-GB10", 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 saricles/MiniMax-M2.7-NVFP4-GB10 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "saricles/MiniMax-M2.7-NVFP4-GB10"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "saricles/MiniMax-M2.7-NVFP4-GB10",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/saricles/MiniMax-M2.7-NVFP4-GB10

SGLang

How to use saricles/MiniMax-M2.7-NVFP4-GB10 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 "saricles/MiniMax-M2.7-NVFP4-GB10" \
    --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": "saricles/MiniMax-M2.7-NVFP4-GB10",
		"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 "saricles/MiniMax-M2.7-NVFP4-GB10" \
        --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": "saricles/MiniMax-M2.7-NVFP4-GB10",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use saricles/MiniMax-M2.7-NVFP4-GB10 with Docker Model Runner:
```
docker model run hf.co/saricles/MiniMax-M2.7-NVFP4-GB10
```

Corrupted / truncated files?

by rrauenza - opened Jun 22

Discussion

rrauenza

Jun 22

•

edited Jun 22

I don't know if it's just the CDN, but three shards download short of their reported bytes. I've tried with curl, with and without xet:

Shard	HF Metadata (correct)	CDN serves (corrupt)	Short by
model-00017-of-00027.safetensors	5,033,633,825 bytes	4,998,786,780 bytes	33.2 MB
model-00018-of-00027.safetensors	5,034,950,770 bytes	5,000,851,964 bytes	32.4 MB
model-00019-of-00027.safetensors	5,011,011,051 bytes	4,998,613,792 bytes	11.8 MB

for i in 17 18 19; do
  shard=\"model-000\${i}-of-00027.safetensors\"
  # Follow redirect fully and check Content-Length at the CDN origin
  cdn_size=\$(curl -sI -L \"https://huggingface.co/saricles/MiniMax-M2.7-NVFP4-GB10/resolve/main/\$shard\" 2>/dev/null | grep -i '^content-length:' | tail -1 | awk '{print \$2}' | tr -d '\r')
  expected=\$(curl -sI \"https://huggingface.co/saricles/MiniMax-M2.7-NVFP4-GB10/resolve/main/\$shard\" 2>/dev/null | grep -i 'x-linked-size' | awk '{print \$2}' | tr -d '\r')
  echo \"\$shard: CDN says \$cdn_size bytes, HF metadata says \$expected bytes\"
done
" --become --become-user=root --limit spark1.private.shroop.net -u ansible 2>&1 | tail -8
spark1.private.shroop.net | CHANGED | rc=0 >>
model-00017-of-00027.safetensors: CDN says 4998786780 bytes, HF metadata says X-Repo-Commit,X-Request-Id,X-Error-Code,X-Error-Message,X-Total-Count,ETag,Link,Accept-Ranges,Content-Range,X-Linked-Size,X-Linked-ETag,X-Xet-Hash
5033633825 bytes
model-00018-of-00027.safetensors: CDN says 5000851964 bytes, HF metadata says X-Repo-Commit,X-Request-Id,X-Error-Code,X-Error-Message,X-Total-Count,ETag,Link,Accept-Ranges,Content-Range,X-Linked-Size,X-Linked-ETag,X-Xet-Hash
5034950770 bytes
model-00019-of-00027.safetensors: CDN says 4998613792 bytes, HF metadata says X-Repo-Commit,X-Request-Id,X-Error-Code,X-Error-Message,X-Total-Count,ETag,Link,Accept-Ranges,Content-Range,X-Linked-Size,X-Linked-ETag,X-Xet-Hash
5011011051 bytes

rrauenza

Jun 22

•

edited Jun 22

/.MEDIA/tmp/MiniMax-M2.7-NVFP4-GB10$ git lfs pull
Skipped checkout for "model-00018-of-00027.safetensors", content not local. Use fetch to download.
Skipped checkout for "model-00019-of-00027.safetensors", content not local. Use fetch to download.
Skipped checkout for "model-00017-of-00027.safetensors", content not local. Use fetch to download.
Downloading LFS objects: 28% (8/29), 56 GB | 76 MB/s

So there's something wrong with the lfs as well. (although I think it just pulls them from cloudfront as well.)

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