| MiniMax H3 is licensed under the MiniMax H3 Community License Agreement, Copyright © 2026 MiniMax. All Rights Reserved. | |
| The files in this repository are modified (quantized) derivatives of MiniMax-H3 | |
| (https://huggingface.co/MiniMaxAI/MiniMax-H3): | |
| - The transformer's per-block AdaLN projections are replaced by an equivalent rank-12 factorization | |
| (ComfyUI curve form: `adaln_t_table` plus per-block projections). | |
| - Transformer GEMM weights are quantized to NVFP4 and/or FP8 with NVIDIA Model Optimizer, and keys are | |
| renamed and fused into ComfyUI's layout. | |
| TaoMate-H3 streaming inference modifications by the Alibaba TaoLive AIGC Team, 2026. | |
| `distill_loras/taomate_h3_fl2va_step300_bf16.safetensors` is a modified copy of TaoMate-H3-FL2VA | |
| (https://huggingface.co/TaoLiveAIGC/TaoMate-H3-FL2VA): keys renamed into ComfyUI's LoRA layout and weights | |
| rounded from fp32 to bf16. | |
| The MiniMax-H3 text encoder is based on Qwen3-VL-32B, licensed under the Apache License 2.0. | |