MiniMax H3 INT4 量化与混合精度 VAE 模型发布
原标题:koongrizzly/MiniMax_H3_int4_W4A8_ConvRot_Pruned
AI 摘要
Hugging Face 用户 koongrizzly 发布了 MiniMax H3 的量化模型合集,包含来自 Winnougan 的 INT4/W4A8 ConvRot 量化 transformer 与文本编码器,以及自行转换的混合 FP16/FP32 Video VAE,并整合了 Turbo LoRA。该仓库面向消费级 GPU 的本地低显存推理,官方适配 ComfyUI,同时提供独立应用仓库。
正文节选
--- license: apache-2.0 tags: - minimax - minimax-h3 - video-generation - audio-video-generation - int4 - convrot - w4a8 - quantization --- # Files for the MiniMax H3 standalone app : https://github.com/Koongrizzly/MiniMax_H3_Standalone_app (also allows loading some int8 models) For questions about the app or repo : https://discord.com/channels/1254516511486513202/1254516512296009882 # MiniMax H3 INT4 ConvRot + Mixed FP16/FP32 Video VAE This repository contains a collection of MiniMax H3 m