Image-to-Video
MiniMax H3
comfyui
lora
text-to-video
video-generation
audio-generation
diffusion
four-step
Instructions to use joyfox/MiniMax-H3-Turbo with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Inference
- Notebooks
- Google Colab
- Kaggle
Load minimax_h3_fl2va_pruned_bf16.safetensors as the diffusion model?
#5
by FanFan2 - opened
In ComfyUI usage, you said "Load minimax_h3_fl2va_pruned_bf16.safetensors as the diffusion model", but the shape of "blocks.25.adaln_proj.linear.weight" is wrong. It seems MiniMax-H3-FL2VA-pruned_rank8_bf16.safetensors is right. So which weight is right ?
https://huggingface.co/DeepBeepMeep/MiniMax-H3/blob/main/MiniMax-H3-FL2VA-pruned_rank8_bf16.safetensors
https://huggingface.co/DeepBeepMeep/MiniMax-H3/blob/main/MiniMax-H3-FL2VA-pruned_bf16.safetensors