Tag: model quantization
Browse our exclusive articles!
3DTurboQuant: Training-Free Quantization for 3D Models
3DTurboQuant enables near-optimal, training-free quantization for 3D reconstruction models, boosting compression with minimal quality loss.
Efficient Multi-LoRA Generative Vision Models on Edge
Discover QUAD: a unified quantization method enabling multi-LoRA generative vision models with reduced memory and faster performance on edge devices.
OneComp: Simplifying Generative AI Model Compression
Discover OneComp, the open-source tool that streamlines generative AI model compression with mixed-precision and progressive quantization.