Instructions to use ISTA-DASLab/Kimi-K2.5-P48NVFP4-MoESQ with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use ISTA-DASLab/Kimi-K2.5-P48NVFP4-MoESQ with Transformers:

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

pipe = pipeline("text-generation", model="ISTA-DASLab/Kimi-K2.5-P48NVFP4-MoESQ", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

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

processor = AutoProcessor.from_pretrained("ISTA-DASLab/Kimi-K2.5-P48NVFP4-MoESQ", trust_remote_code=True)
model = AutoModel.from_pretrained("ISTA-DASLab/Kimi-K2.5-P48NVFP4-MoESQ", trust_remote_code=True, device_map="auto")
messages = [
    {"role": "user", "content": "Who are you?"},
]
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 ISTA-DASLab/Kimi-K2.5-P48NVFP4-MoESQ with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "ISTA-DASLab/Kimi-K2.5-P48NVFP4-MoESQ"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ISTA-DASLab/Kimi-K2.5-P48NVFP4-MoESQ",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/ISTA-DASLab/Kimi-K2.5-P48NVFP4-MoESQ

SGLang

How to use ISTA-DASLab/Kimi-K2.5-P48NVFP4-MoESQ 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 "ISTA-DASLab/Kimi-K2.5-P48NVFP4-MoESQ" \
    --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": "ISTA-DASLab/Kimi-K2.5-P48NVFP4-MoESQ",
		"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 "ISTA-DASLab/Kimi-K2.5-P48NVFP4-MoESQ" \
        --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": "ISTA-DASLab/Kimi-K2.5-P48NVFP4-MoESQ",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use ISTA-DASLab/Kimi-K2.5-P48NVFP4-MoESQ with Docker Model Runner:
```
docker model run hf.co/ISTA-DASLab/Kimi-K2.5-P48NVFP4-MoESQ
```

diamond degradation

by MB66166 - opened Sep 30

Discussion

MB66166

Sep 30

It's huge, 20 point loss on diamond level reasoning

maybe RCO , not MoESQ, and if possible a shard 2 in SDD?

MB66166

Sep 30

Sorry, "only" almost 15, but still too big, we're talking about ternary degradation level

BTW as a general rule, thnks for all efforts, but this road seems less promising.

kwanhee

IST Austria Distributed Algorithms and Systems Lab org 27 days ago

thank you for your interest!

inducing sparsity constraint is usually harder than lower bitwidth; maybe you can check out our new K3-P48NVFP4, which has lossless compression on reasoning benchmarks.

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