Repository Issue Activity (beta)

#1672 目前h3的int8模型是不支持多卡训练吗

#1 Upload examples via this issue

#1557 MiniMax-H3这个模型经过了CFG-Distillation，它还能正常SFT吗？会不会SFT更多step以后丧失CFG蒸馏的能力？

#1663 train_scale_cfg

#1669 controlnet_aux is required by the Annotator examples but is not in any install extra

#1668 Partial download surfaces as "Cannot detect the model type" rather than a download error

#1664 LingBot-Video 30B LoRA training OOM on H100

#1625 可否支持 SenseNova-U1 系列

#1660 minimax-H3 FL2VA train error

#1617 DMD2训练损失震荡

#1645 Minimax h3 多卡如何推理？

#1609 [Proposal] 探讨对 Diffusion Language Model 的支持

#1639 minimax h3我看新增了 adater 训练，ref2av 目前是不支持这种训练方式吗

#1642 训练minimax- h3的nf4模型的lora，使用adapter训练时融合lora会报错

#994 Question about optimal hyper parameters for training Qwen-Image-Edit-2509

#1605 Consider using pre-built Flash Attention kernels (and others) via kernels

#1585 the training details of minimax h3 model

#1324 FLUX.2 dev generations mismatch with diffusers

#1624 Question: Minimum GPU Memory Required for Training MiniMax H3 Video LoRA

#1629 Support MinMax-H3 training and inference

#1626 Weird results when editing images with HiDream-O1-Image-Dev

#1588 会考虑支持d-opsd吗

#1587 训练minimax-h3的lora没办法在comfyui中使用

#1613 --processor_path 传本地路径时丢失 path，导致 processor 加载报 "No valid model files"

#1600 Bug report: disk_map.py 的 device == "cpu" 失效导致权重未复制 → 进程崩溃 + 内存膨胀