Commit History
-
Point component loading specs at this repository
Verified signature82cfd00 -
Add diffusers format weights
Verified signatureAdds the diffusers format weights at the root of the repository, alongside the original checkpoint folders. The layout is component wise, so diffusers loading fetches exactly the subfolders a task needs and never touches FL2VA/ or Ref2VA/. transformer/ serves text to video+audio and first/last frame conditioning, transformer_ref/ serves omni reference conditioning, and text_encoder/, tokenizer/, processor/, vae/, audio_vae/, scheduler/ and audio_scheduler/ are shared by both. modular_model_index.json is the loading manifest. Requires the diffusers integration: https://github.com/huggingface/diffusers/pull/14355 (usage and per hardware guidance: https://github.com/huggingface/diffusers/blob/minimax-h3/docs/source/en/api/pipelines/minimax_h3.md) ```python import torch from diffusers import ModularPipeline pipe = ModularPipeline.from_pretrained("MiniMaxAI/MiniMax-H3") pipe.load_components(dtype=torch.bfloat16) ``` The converted transformer reproduces the reference implementation's de…4ae026f -
Init MiniMaxAI/MiniMax-H3
Co-authored-by: yuanhe <yuanhe@users.noreply.huggingface.co> Co-authored-by: RuggingHace <RuggingHace@users.noreply.huggingface.co> Co-authored-by: MiniMax-AI <MiniMax-AI@users.noreply.huggingface.co>af0fe5a
