modelscope download --dataset DiffSynth-Studio/diffsynth_example_dataset --include "minimax_h3/MiniMax-H3-Ref2VA/*" --local_dir ./data/diffsynth_example_dataset # Optional: fuse the DeCFG training adapter into the DiT while training, for a better optimization landscape on this CFG-distilled base. This DiT uses the ComfyUI qkv layout, so take the model_for_comfy_dit variant. Training only -- do not load it at inference. # modelscope download --model DiffSynth-Studio/MiniMax-H3-TrainingAdapter --include model_ref2va_for_comfy_dit.safetensors --local_dir ./models/DiffSynth-Studio/MiniMax-H3-TrainingAdapter # --preset_lora_path "./models/DiffSynth-Studio/MiniMax-H3-TrainingAdapter/model_ref2va_for_comfy_dit.safetensors" \ # --preset_lora_model "dit" # Ref2VA - stage 1: encode reference/text/video/audio and cache (text encoder + processor loaded here only) accelerate launch examples/minimax_h3/model_training/train.py \ --dataset_base_path data/diffsynth_example_dataset/minimax_h3/MiniMax-H3-Ref2VA \ --dataset_metadata_path data/diffsynth_example_dataset/minimax_h3/MiniMax-H3-Ref2VA/metadata.json \ --data_file_keys "video,input_audio,references" \ --extra_inputs "input_audio,references" \ --height 480 \ --width 832 \ --num_frames 124 \ --dataset_repeat 1 \ --model_id_with_origin_paths "Comfy-Org/MiniMax-H3:text_encoders/qwen3vl_32b_minimax_h3_int8_convrot.safetensors,Comfy-Org/MiniMax-H3:diffusion_models/minimax_h3_ref2va_int8_convrot.safetensors,MiniMax/MiniMax-H3:Ref2VA/video_vae/source/model.safetensors,MiniMax/MiniMax-H3:Ref2VA/audio_vae/model.safetensors" \ --processor_path "MiniMax/MiniMax-H3:Ref2VA/processor/" \ --remove_prefix_in_ckpt "pipe.dit." \ --output_path "./models/train/MiniMax-H3-Ref2VA-int8-convrot-cache" \ --lora_base_model "dit" \ --lora_target_modules "attn.qkv_proj,attn.out_proj,mlp.fc1,mlp.fc2" \ --lora_rank 32 \ --use_gradient_checkpointing \ --task "sft:data_process" # Ref2VA - stage 2: train LoRA on cached inputs (only the DiT is loaded) accelerate launch examples/minimax_h3/model_training/train.py \ --dataset_base_path "./models/train/MiniMax-H3-Ref2VA-int8-convrot-cache" \ --num_frames 124 \ --dataset_repeat 100 \ --model_id_with_origin_paths "Comfy-Org/MiniMax-H3:diffusion_models/minimax_h3_ref2va_int8_convrot.safetensors" \ --learning_rate 1e-4 \ --num_epochs 5 \ --remove_prefix_in_ckpt "pipe.dit." \ --output_path "./models/train/MiniMax-H3-Ref2VA-int8-convrot" \ --lora_base_model "dit" \ --lora_target_modules "attn.qkv_proj,attn.out_proj,mlp.fc1,mlp.fc2" \ --lora_rank 32 \ --use_gradient_checkpointing \ --find_unused_parameters \ --task "sft:train"