You can deploy this model and start earning money today!

cybermotaz/
qwen3-vl-2b-thinking-nvfp4-w4a16

Name: qwen3-vl-2b-thinking-nvfp4-w4a16
Availability: InStock

Views

No views yet

Qwen3-VL-2B-Thinking NVFP4 W4A16

First NVFP4 Quantization of Qwen3-VL-2B-Thinking

By Mutaz Al Awamleh | ELK-AI

Docker

HuggingFace

License

Model Description

This is the first publicly available NVFP4 W4A16 quantized version of Qwen3-VL-2B-Thinking, a vision-language model optimized for NVIDIA Blackwell (SM121) architecture.

Attribute	Original	NVFP4 Quantized
Parameters	2B	Same
Architecture	Vision-Language + Thinking	Same
Model Size	~4.0 GB	~2.1 GB
Memory Savings	-	48%
Precision	BF16	FP4 W4A16

Quick Start

Using vLLM (Recommended)

python

from vllm import LLM, SamplingParams

model = LLM(
    model="cybermotaz/qwen3-vl-2b-thinking-nvfp4-w4a16",
    trust_remote_code=True,
    quantization="modelopt_fp4",
    kv_cache_dtype="fp8",
    gpu_memory_utilization=0.95
)

sampling_params = SamplingParams(temperature=0.7, max_tokens=512)
prompt = "Think step by step: What is shown in this image?"

outputs = model.generate([prompt], sampling_params)
print(outputs[0].outputs[0].text)

Using Docker (Pre-loaded)

bash

# Pull the optimized container
docker pull elkaioptimization/vllm-nvfp4-cuda-13:qwen3-vl-2b-thinking-nvfp4-1.0

# Run with OpenAI-compatible API
docker run --gpus all -p 8000:8000 \
    elkaioptimization/vllm-nvfp4-cuda-13:qwen3-vl-2b-thinking-nvfp4-1.0

Quantization Details

Parameter	Value
Quantization Format	NVFP4 (FP4 E2M1)
Weight Precision	4-bit (W4)
Activation Precision	16-bit (A16)
Block Size	16 elements
Scale Format	FP8 E4M3
Calibration Dataset	CNN/DailyMail (512 samples)
Calibration Method	AWQ-style
Tool Used	NVIDIA TensorRT-Model-Optimizer

Hardware Requirements

Requirement	Minimum	Recommended
GPU	RTX 3060 (6GB)	RTX 4090 / DGX Spark
GPU Memory	6 GB	24 GB+
CUDA	12.4+	13.0
Driver	560+	570+

Model Architecture

Qwen3-VL-2B-Thinking features:

Vision-Language: Processes both images and text inputs
Enhanced Reasoning: Optimized for step-by-step thinking and complex reasoning
Extended Context: 32K native, 262K extended context length
Multilingual: Strong performance in English and Chinese

Links

Resource	Link
Original Model	Qwen/Qwen3-VL-2B-Thinking
Docker (Org)	elkaioptimization/vllm-nvfp4-cuda-13
Docker (Personal)	mutazai/vllm-spark-blackwell-nvfp4-optimized
Author	Mutaz Al Awamleh
Organization	ELK-AI

License

This model is released under the Apache 2.0 License, same as the original Qwen3 model.

Built by Mutaz Al Awamleh | ELK-AI

First to quantize Qwen3-VL-2B-Thinking to NVFP4 for Blackwell