Z-Image-Turbo is an image processing model quantized in FP8 E5M2 and E4M3FN formats. It is optimized based on the original Tongyi-MAI/Z-Image-Turbo model, significantly reducing the model size and inference resource requirements while maintaining performance.

MultimodalDiffusersEnglish

An optimized image generation model based on black-forest-labs/FLUX.2-dev, supporting image-to-image generation and editing tasks. It uses fp8_scaled quantization technology to improve inference speed and is particularly suitable for use on low-VRAM GPUs.

This is an automatic speech recognition model optimized for Apple silicon chip devices. By converting to the MLX framework and quantizing to the FP8 format, it enables fast on-device speech transcription on Apple devices. The model is fine-tuned for verbatim accuracy and is particularly suitable for scenarios requiring high-precision transcription.

Audio ProcessingMlxOther

GLM-4.5-Iceblink-v2-106B-A12B-FP8 is a version based on the GLM-4.5-Iceblink-v2-106B-A12B model, using the most advanced mixture of experts quantization method for FP8 quantization. This model is specifically optimized for Ada, Hopper, or Blackwell series GPUs that support hardware FP8, significantly improving inference efficiency while maintaining high-quality output.

Natural Language ProcessingSafetensors

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

This is the FP8 quantized version of the Meta Llama-3.1-8B-Instruct model. By performing FP8 quantization on weights and activation values, it significantly reduces the disk size and GPU memory requirements while maintaining good model performance.

Natural Language ProcessingSafetensors

The FP8 quantized version of olmOCR-2-7B-1025, fine-tuned based on Qwen2.5-VL-7B-Instruct, is a vision-language model specifically designed for handling complex OCR scenarios such as mathematical equations and tables.

MultimodalTransformersEnglish

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

Granite 4.0 H-Small (FP8) is a small FP8 quantized version of the Granite series of language models developed by IBM, designed to provide professional, accurate, and secure responses. The model uses the FP8 precision format, optimizing inference efficiency while maintaining performance.

Natural Language ProcessingTransformers

This is the FP8 dynamic quantization version of the NVIDIA-Nemotron-Nano-9B-v2 model. Optimization is achieved by quantizing weights and activations to the FP8 data type, significantly reducing disk size and GPU memory requirements by approximately 50% while maintaining excellent text generation performance.

Natural Language ProcessingTransformersMultiple Languages

This is a quantized version of Qwen3-VL-235B-A22B-Instruct. By quantizing the weights and activation values to the FP8 data type, it effectively reduces the disk size and GPU memory requirement by approximately 50%. It supports text, image, and video inputs and outputs text, suitable for various natural language processing and multimodal tasks.

MultimodalSafetensors

Qwen3-VL-235B-A22B-Instruct-FP8-dynamic is a quantized version of Qwen3-VL-235B-A22B-Instruct. By quantizing weights and activations to the FP8 data type, it significantly reduces disk size and GPU memory requirements while maintaining high model performance.

MultimodalSafetensors

This is the FP8 quantized version of the Qwen3-Next-80B-A3B-Instruct model, supported by the vLLM framework. While maintaining the model quality, it significantly reduces the model size and improves the inference efficiency.

Natural Language ProcessingSafetensors

This is the version after NVIDIA quantizes Alibaba's Qwen3-14B model to FP8. It adopts an optimized Transformer architecture, supports a context length of 131K, and is suitable for various AI application scenarios.

Natural Language ProcessingSafetensors

NVIDIA Qwen3-8B FP8 is a quantized version of the Alibaba Qwen3-8B model. It uses an optimized Transformer architecture and belongs to the autoregressive language model. This model is optimized through FP8 quantization technology and can achieve efficient inference on NVIDIA GPUs. It supports both commercial and non-commercial use.

Natural Language ProcessingSafetensors