Instructions to use vistralis/Qwen3-4B-FP8 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use vistralis/Qwen3-4B-FP8 with Transformers:

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

pipe = pipeline("text-generation", model="vistralis/Qwen3-4B-FP8")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("vistralis/Qwen3-4B-FP8")
model = AutoModelForCausalLM.from_pretrained("vistralis/Qwen3-4B-FP8")
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 vistralis/Qwen3-4B-FP8 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "vistralis/Qwen3-4B-FP8"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "vistralis/Qwen3-4B-FP8",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/vistralis/Qwen3-4B-FP8

SGLang

How to use vistralis/Qwen3-4B-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 "vistralis/Qwen3-4B-FP8" \
    --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": "vistralis/Qwen3-4B-FP8",
		"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 "vistralis/Qwen3-4B-FP8" \
        --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": "vistralis/Qwen3-4B-FP8",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use vistralis/Qwen3-4B-FP8 with Docker Model Runner:
```
docker model run hf.co/vistralis/Qwen3-4B-FP8
```

Qwen3-4B-FP8

FP8 (W8A8) quantized version of Qwen/Qwen3-4B, created using llm-compressor with calibrated quantization.

Overview

Property	Value
Base Model	Qwen/Qwen3-4B
Parameters	4.41B
Quantization	FP8 (W8A8)
Format	`compressed-tensors`
Tool	llm-compressor
Disk Size	~4.9 GB (2 shards)
VRAM	~3.96 GB

Intended Use

Quantized text encoder for Flux 2 Klein 4B image generation pipelines. Architecturally identical to the Klein 4B text encoder.

Quantization Details

Scheme: FP8 — 8-bit float weights and activations (float8_e4m3fn)
Targets: All Linear layers (excluding lm_head)
Calibration: 256 samples, sequential pipeline with CPU offloading

Hardware Requirements

Minimum: NVIDIA Hopper (CC 8.9+) or Ada Lovelace for native FP8 inference
Fallback: Dequantizes to BF16 on older hardware

Downloads last month: 12

Safetensors

Model size

4B params

Tensor type

BF16

F8_E4M3

Model tree for vistralis/Qwen3-4B-FP8

Base model

Qwen/Qwen3-4B-Base

Finetuned

Qwen/Qwen3-4B

Quantized

(218)

this model