search

ShivamShrirao / diffusers

🤗 Diffusers: State-of-the-art diffusion models for image and audio generation in PyTorch
https://huggingface.co/docs/diffusers
Apache License 2.0
1.89k stars 505 forks source link

remove keep_fp32_wrapper #181

Closed KitaharaMugiro closed 1 year ago

ShivamShrirao commented 1 year ago

You need to update your accelerate version for it to work.

alex-pub commented 1 year ago

At the moment, https://github.com/ShivamShrirao/diffusers/blob/main/examples/dreambooth/requirements.txt#L1 has Accelerate version pinned to 0.12.0 which doesn't have the keep_fp32_wrapper parameter in its Accelerate.accelerator.unwrap_model signature, so either requirements.txt needs an update, or the save_weights changes need to be reverted to bring back train_dreambooth.py into a working state.

ShivamShrirao commented 1 year ago

@alex-pub thanks. Updated requirements.txt