This is an enhanced multimodal vision-language model based on the Qwen3-VL-8B-Thinking model. It is extended to 12B parameters through Brainstorm 20x technology and uses NEO Imatrix-enhanced GGUF quantization. The model has powerful image understanding, text generation, and multimodal reasoning capabilities, with significant improvements in visual perception, text quality, and creative scenarios.

MultimodalGgufMultiple Languages

NVIDIA-Nemotron-Nano-VL-12B-V2-FP4-QAD is an autoregressive vision-language model launched by NVIDIA. Based on an optimized Transformer architecture, it can handle both image and text inputs simultaneously. The model uses FP4 quantization technology, which significantly reduces the model size and inference cost while maintaining performance, and is suitable for various multimodal application scenarios.

MultimodalTransformers

Qwen3-VL-32B-Thinking-FP8 is the FP8 quantized version of the most powerful vision-language model in the Qwen series. It uses fine-grained fp8 quantization technology with a block size of 128, and its performance indicators are almost the same as those of the original BF16 model. This model has excellent text understanding and generation capabilities, in-depth visual perception and reasoning abilities, long context support, and powerful agent interaction capabilities.

MultimodalSafetensors

This is a vision Transformer model based on the DINOv3 framework, trained on the LVD-1689M dataset from the DINOv3 ViT-7B model through knowledge distillation technology. This model is specifically designed for image feature encoding and can efficiently extract image feature representations, suitable for various computer vision tasks.

Computer VisionTimm

This is a computer vision model fine-tuned based on Ultralytics/YOLO11, specifically designed to detect UI elements in macOS application screenshots. It is part of the Screen2AX project, dedicated to generating accessibility metadata using computer vision technology.

fastvlm-0.5b-captions is an efficient vision-language model fine-tuned based on the third-stage model of FastVLM-0.5B, designed specifically for structured image description on mobile devices. This model uses LoRA fine-tuning, 4-bit quantization, and MobileCLIP-S0 visual tower technology, significantly reducing memory usage and enabling real-time inference on mobile devices such as iPhones.

MultimodalTransformersEnglish

A zero-shot skin disease assessment model based on vision-language pre-training technology, integrating multi-faceted knowledge enhancement to provide an effective tool for skin disease research and diagnosis.

MultimodalOpenclip

OpenVLA-OFT is a vision-language-action model optimized based on the OpenVLA model. Through fine-tuning technology, it significantly improves the execution speed and success rate on the LIBERO task suite.

MultimodalTransformers

Ristretto is an innovative vision-language model that employs dynamic image token deployment technology, allowing flexible adjustment of image token quantities based on task requirements, surpassing previous generations in performance and versatility.

MultimodalTransformersMultiple Languages

Llama 3.2 11B vision instruction fine-tuned model optimized with Unsloth, using 4-bit quantization technology, achieving 2x faster training speed

MultimodalTransformersEnglish

OpenVLA-OFT is an optimized Vision-Language-Action model designed for robot tasks, which improves speed and success rate through fine-tuning technology.

MultimodalTransformers

OpenVLA-OFT is an optimized vision-language-action model that significantly improves the performance and speed of the basic OpenVLA model through fine-tuning technology.

MultimodalTransformers

OpenVLA-OFT is an optimized vision-language-action model that significantly improves the running speed and task success rate of the basic OpenVLA model through fine-tuning technology.

MultimodalTransformers

Vision Transformer model based on SigLIP 2 technology, focused on image feature extraction

MultimodalTransformers

A compact vision Transformer model trained with DINOv2, utilizing register technology to optimize attention mechanisms and enhance image feature extraction capabilities

Computer VisionTransformers

This is an OpenVLA 7B vision-language-action model fine-tuned using LoRA technology on the LIBERO-Goal dataset, suitable for the robotics field.

MultimodalTransformersEnglish

Realistic Vision v1.3 is a text-to-image generation model based on stable diffusion technology, specializing in hyper-realistic, realistic, and semi-realistic style images. The model supports 768x768 resolution and emphasizes compositional freedom and creativity.

Computer VisionDiffusers

The Natural Vision Lightning version is a lightweight and fast version based on the Natural Vision model. It combines SDXL Lightning technology and can quickly generate high-quality images at different steps, suitable for various image generation scenarios.

Computer VisionDiffusersEnglish

The CVLFace pre-trained model is a face recognition model based on the Vision Transformer architecture. It uses the Keypoint Relative Position Encoding (KPRPE) technology and is trained on the WebFace4M dataset, providing powerful feature extraction capabilities for face recognition tasks.

Computer VisionTransformersEnglish

CheXagent is a foundation model focused on chest X-ray interpretation, aiming to enhance the accuracy and efficiency of medical imaging analysis through advanced vision-language processing technology.

Computer VisionTransformers