Please add a cell in dreambooth notebook to stop the image generator cause of OOM

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

1.88k stars 509 forks source link

Please add a cell in dreambooth notebook to stop the image generator cause of OOM #12

Open 1blackbar opened 1 year ago

1blackbar commented 1 year ago

I tested the model and i wanted to retrain new one, i cant ccause model is loaded into memory so please add a cell to unload it so we can train again instesad of disconnecting and running it over from 0 It shows out of memory when i run training after i generated imgs

brentjohnston commented 1 year ago

Is that why I can't find the source ckpt file in VisComp (wherever that folder is)? I wanted to try to train a custom model I had but I couldn't find the model anywhere. Also I notice if you disable the hugginface token line you can't train.

pink-red commented 1 year ago

You can restart the runtime (Runtime -> Restart runtime), this will preserve the installed libraries and files. After this, start from the "Login to HuggingFace" cell.

DON'T do "Disconnect and delete runtime" for that, you will start from scratch.