Instructions to use liodon-ai/Nanbeige4.2-3B-DSpark-FP8 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use liodon-ai/Nanbeige4.2-3B-DSpark-FP8 with Transformers:

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

pipe = pipeline("text-generation", model="liodon-ai/Nanbeige4.2-3B-DSpark-FP8", trust_remote_code=True)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("liodon-ai/Nanbeige4.2-3B-DSpark-FP8", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("liodon-ai/Nanbeige4.2-3B-DSpark-FP8", trust_remote_code=True, device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use liodon-ai/Nanbeige4.2-3B-DSpark-FP8 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "liodon-ai/Nanbeige4.2-3B-DSpark-FP8"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "liodon-ai/Nanbeige4.2-3B-DSpark-FP8",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/liodon-ai/Nanbeige4.2-3B-DSpark-FP8

SGLang

How to use liodon-ai/Nanbeige4.2-3B-DSpark-FP8 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 "liodon-ai/Nanbeige4.2-3B-DSpark-FP8" \
    --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": "liodon-ai/Nanbeige4.2-3B-DSpark-FP8",
		"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 "liodon-ai/Nanbeige4.2-3B-DSpark-FP8" \
        --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": "liodon-ai/Nanbeige4.2-3B-DSpark-FP8",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use liodon-ai/Nanbeige4.2-3B-DSpark-FP8 with Docker Model Runner:
```
docker model run hf.co/liodon-ai/Nanbeige4.2-3B-DSpark-FP8
```

Nanbeige4.2-3B-DSpark-FP8 / README.md

PY-AI-Dev

Add FP8 (dynamic) quantization for Nanbeige4.2-3B-DSpark

8bd13df verified 5 days ago

preview code

Raw

History Blame Contribute Delete

2.17 kB

metadata

license: other
base_model: Nanbeige/Nanbeige4.2-3B-DSpark
base_model_relation: quantized
library_name: transformers
pipeline_tag: text-generation
tags:
  - fp8
  - compressed-tensors
  - vllm
  - quantized
quantized_by: liodon-ai

Nanbeige4.2-3B-DSpark — FP8 (dynamic)

FP8 quantization of Nanbeige/Nanbeige4.2-3B-DSpark, published by Liodon AI.

Quantized with llm-compressor using the FP8_DYNAMIC scheme: weights are cast to FP8 (E4M3) per-channel ahead of time, activations are quantized to FP8 dynamically per-token at inference time. No calibration dataset is needed for this scheme, so the quantized weights are numerically just a direct cast of the original — no calibration-set bias to worry about. lm_head is left unquantized (standard practice — negligible size, disproportionate quality impact if quantized).

Original size: 1.7 GB → Quantized: 2.8 GB.

Quick Start

vLLM

vllm serve liodon-ai/Nanbeige4.2-3B-DSpark-FP8

Text Generation Inference (TGI)

docker run --gpus all -p 8080:80 ghcr.io/huggingface/text-generation-inference \
    --model-id liodon-ai/Nanbeige4.2-3B-DSpark-FP8

SGLang

python -m sglang.launch_server --model-path liodon-ai/Nanbeige4.2-3B-DSpark-FP8

FP8 execution requires an NVIDIA GPU with compute capability ≥ 8.9 (Ada/Hopper/Blackwell — RTX 40-series, L4/L40S, H100/H200, B100/B200/GB10). On older GPUs, vLLM/TGI will dequantize to run, which loses the speed/memory benefit.

Source

Model: Nanbeige/Nanbeige4.2-3B-DSpark
License: other

Citation

@misc{liodonai_nanbeige4_2_3b_dspark_fp8,
  title        = {Nanbeige4.2-3B-DSpark — FP8},
  author       = {{Liodon AI}},
  year         = {2026},
  howpublished = {\url{https://huggingface.co/liodon-ai/Nanbeige4.2-3B-DSpark-FP8}},
  note         = {FP8 (dynamic) quantization of Nanbeige/Nanbeige4.2-3B-DSpark}
}

Quantized by Liodon AI