This is a version of the Qwen3-Next-80B large language model after 4-bit quantization using Huawei's SINQ (Sinkhorn Normalized Quantization) method. SINQ is a novel, fast, and high-quality quantization method designed to significantly compress the model size (by approximately 75%) while maintaining almost the same accuracy as the original model, making it easier to deploy.

Natural Language ProcessingSafetensorsEnglish

Qwen3-VL-2B-Instruct-FP8 is the FP8 quantized version of the most powerful vision-language model in the Qwen series. It uses fine-grained fp8 quantization with a block size of 128, and its performance is almost the same as that 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 enhanced spatial and video dynamic understanding abilities.

MultimodalTransformers

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

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

Qwen3-VL-30B-A3B-Thinking-FP8 is the FP8 quantized version of the most powerful vision-language model in the Tongyi series. It uses fine-grained fp8 quantization 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, and supports long contexts and video understanding.

MultimodalSafetensors

Qwen3-VL-30B-A3B-Instruct-FP8 is the FP8 quantized version of the most powerful vision-language model in the Tongyi series. It uses fine-grained FP8 quantization with a block size of 128, and its performance is almost the same as that of the original BF16 model. This model has excellent text understanding and generation capabilities, in-depth visual perception and reasoning capabilities, long context support, and strong agent interaction capabilities.

MultimodalSafetensors

Qwen3-VL-235B-A22B-Thinking-FP8 is the FP8 quantized version of the most powerful vision-language model in the Tongyi series. It uses fine-grained fp8 quantization with a block size of 128, and its performance metrics 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 capabilities, long context support, and strong agent interaction capabilities.

MultimodalSafetensors

Qwen3-VL-235B-A22B-Instruct-FP8 is the FP8 quantized version of the most powerful vision-language model in the Tongyi series. It uses fine-grained fp8 quantization with a block size of 128, and its performance is almost the same as that 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 strong spatial and video dynamic understanding abilities.

MultimodalSafetensors

This project is the A3B quantized version of Alibaba's Tongyi In-depth Research 30B model. It aims to reduce the model deployment cost through quantization technology and make knowledge accessible to everyone. This model is optimized based on a large language model with a scale of 30B parameters, maintaining the strong capabilities of the original model while improving the inference efficiency.

Natural Language ProcessingGguf

This is a ViT-L14 image encoder based on the PE-Core model by Bolya et al., which has been converted to the Birder format for image feature extraction. The model retains the original weights and architecture but removes the CLIP projection layer to output raw image embeddings. It is a general-purpose visual backbone network suitable for image classification and detection tasks.

Computer VisionPytorch

This is the 3-bit quantization version of the Qwen3-235B-A22B-Thinking-2507 model, optimized and converted for the MLX framework. The original model is a large language model developed by the Alibaba Qwen team, with a scale of 235 billion parameters and support for chain-of-thought reasoning ability.

Natural Language ProcessingMlx

A LoRA trained on the Wan2.1 14B I2V 480p model, capable of achieving smooth 360-degree orbiting camera movements around subjects, providing dynamic all-around perspective showcases.

MultimodalDiffusersEnglish

Orkhon-TTS is a Turkish text-to-speech model based on the F5 TTS architecture, developed by Hasan Can Solakoğlu and currently in the Alpha stage.

Lumiere Alpha is a text-to-image model focused on enhancing realism while maintaining prompt coherence and the original composition of the FLUX.1-Dev model.

Computer VisionDiffusers

This is the 4-bit quantized version of Qwen2-Audio-7B-Instruct, developed based on Alibaba Cloud's original Qwen model. It is an audio-text multimodal large language model.

MultimodalTransformers

CC BY-NC 4.0 is a widely used Creative Commons license that allows users to freely share, use, and adapt works on the condition that the original author is credited and the works are not used for commercial purposes.

Natural Language ProcessingTransformers

This is an open-source project based on the MIT License. Specific model information is not provided in the original document. This model follows the principle of open source code, allowing users to freely use, modify, and distribute it.

Natural Language ProcessingTransformers

Zephyr 141B-A39B is a large language model fine-tuned from Mixtral-8x22B-v0.1, trained using the ORPO alignment algorithm, designed to be a helpful assistant.

Natural Language ProcessingTransformers

This is a dataset of DPO (Direct Preference Optimization) training pairs built on the Intel Orca dataset, specifically designed for preference learning and model alignment training.

Natural Language ProcessingTransformers

The Creative Commons Attribution 4.0 International License (CC BY 4.0) is a license that allows users to freely share, use, and adapt works, provided that the original author is credited.

Natural Language ProcessingTransformers