notebooks
notebooks copied to clipboard
add accelerate notebook_launcher to sd-xl dreambooth
adds accelerate notebook_launcher to improve user experience + some minor ui improvements building on top of #424 training script was copied from train_dreambooth_lora_sdxl.py
cc @sayakpaul @apolinario 🤗
Check out this pull request onÂ
See visual diffs & provide feedback on Jupyter Notebooks.
Powered by ReviewNB
View / edit / reply to this conversation on ReviewNB
sayakpaul commented on 2023-08-16T07:41:26Z ----------------------------------------------------------------
Good idea to link to your DreamBooth dataset here with an example of how to load a particular configuration from there. I think this will encourage folks to experiment more.
View / edit / reply to this conversation on ReviewNB
sayakpaul commented on 2023-08-16T07:41:27Z ----------------------------------------------------------------
Did you clear the output?
View / edit / reply to this conversation on ReviewNB
sayakpaul commented on 2023-08-16T07:41:27Z ----------------------------------------------------------------
I think it's totally okay to:
- Chunk the long code cells into multiple blocks. Since this is a notebook environment, I find long code cells to be very difficult to navigate through.
- Remove a few CLI args and keep them fixed.
A good example of how this could be done can be found here: https://github.com/huggingface/notebooks/blob/main/diffusers/training_example.ipynb
linoytsaban commented on 2023-08-16T07:48:40Z ----------------------------------------------------------------
Agree, will make the changes :). Also, I wanted to keep this cell (training loop) and the args documentation cell closed by default - would be better in this case no? (even after chunking into multiple cell that is),
Agree, will make the changes :). Also, I wanted to keep this cell (training loop) and the args documentation cell closed by default - would be better in this case no? (even after chunking into multiple cell that is),
View entire conversation on ReviewNB