nicollegah
nicollegah
I would also really love this to be implemented.
I'd love a solution for this too!
This would be great! +1
I'm getting `RuntimeError: "replication_pad3d_cuda" not implemented for 'BFloat16'` edit: solved with ``` pip install --upgrade torch pip install --upgrade xformers ```
looks similar for me https://github.com/user-attachments/assets/b6ab9e80-755c-47ab-bcb0-380153fe73f8
> Try using this checkpoint for the transformer https://huggingface.co/imnotednamode/mochi-1-preview-mix-nf4 in the pipeline , its nf4 quantized mixed in with bf16 gives decent results @tin2tin There, it says that we must...
``` import torch from diffusers import MochiPipeline, MochiTransformer3DModel from diffusers.utils import export_to_video pipe = MochiPipeline.from_pretrained("genmo/mochi-1-preview", revision="refs/pr/18", torch_dtype=torch.bfloat16 ) pipe.enable_model_cpu_offload() pipe.enable_vae_tiling() prompt="A panda, dressed in a small, red jacket and a...