Instructions to use RedHatAI/MiniMax-M2.5-quantized.w4a16 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use RedHatAI/MiniMax-M2.5-quantized.w4a16 with Transformers:

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

pipe = pipeline("text-generation", model="RedHatAI/MiniMax-M2.5-quantized.w4a16", 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("RedHatAI/MiniMax-M2.5-quantized.w4a16", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("RedHatAI/MiniMax-M2.5-quantized.w4a16", trust_remote_code=True)
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

vLLM

How to use RedHatAI/MiniMax-M2.5-quantized.w4a16 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "RedHatAI/MiniMax-M2.5-quantized.w4a16"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "RedHatAI/MiniMax-M2.5-quantized.w4a16",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/RedHatAI/MiniMax-M2.5-quantized.w4a16

SGLang

How to use RedHatAI/MiniMax-M2.5-quantized.w4a16 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 "RedHatAI/MiniMax-M2.5-quantized.w4a16" \
    --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": "RedHatAI/MiniMax-M2.5-quantized.w4a16",
		"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 "RedHatAI/MiniMax-M2.5-quantized.w4a16" \
        --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": "RedHatAI/MiniMax-M2.5-quantized.w4a16",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use RedHatAI/MiniMax-M2.5-quantized.w4a16 with Docker Model Runner:
```
docker model run hf.co/RedHatAI/MiniMax-M2.5-quantized.w4a16
```

alexmarques commited on about 24 hours ago

Commit

7c716cd

verified ·

1 Parent(s): 5dbfe8b

Update README.md

Browse files

Files changed (1) hide show

README.md +2 -2

README.md CHANGED Viewed

@@ -88,7 +88,7 @@ from transformers import AutoModelForCausalLM, AutoTokenizer, AutoProcessor
 from llmcompressor import oneshot
 from llmcompressor.modifiers.quantization import GPTQModifier
-MODEL_ID = "inference-optimization/MiniMax-M2.5-BF16"
 # Load model.
 model = AutoModelForCausalLM.from_pretrained(MODEL_ID, torch_dtype="auto", trust_remote_code=True)
@@ -209,7 +209,7 @@ The model was evaluated on the ifeval, mmlu_pro and gsm8k_platinum  using [lm-ev
 ### Accuracy
-| Benchmark | inference-optimization/MiniMax-M2.5-BF16 | inference-optimization/MiniMax-M2.5.w4a16 | Recovery (%) |
 |-----------|------------------------------------------|-------------------------------------------|--------------|
 | GSM8k Platinum (0-shot) | 95.15 | 96.36 | 101.27 |
 | IfEval (0-shot) | 92.05 | 90.45 | 98.26 |

 from llmcompressor import oneshot
 from llmcompressor.modifiers.quantization import GPTQModifier
+MODEL_ID = "RedHatAI/MiniMax-M2.5-BF16"
 # Load model.
 model = AutoModelForCausalLM.from_pretrained(MODEL_ID, torch_dtype="auto", trust_remote_code=True)
 ### Accuracy
+| Benchmark | RedHatAI/MiniMax-M2.5-BF16 | RedHatAI/MiniMax-M2.5.w4a16 | Recovery (%) |
 |-----------|------------------------------------------|-------------------------------------------|--------------|
 | GSM8k Platinum (0-shot) | 95.15 | 96.36 | 101.27 |
 | IfEval (0-shot) | 92.05 | 90.45 | 98.26 |