lllyasviel/ControlNet

Training still OOM on 8GB gpu.

Open

#21 aperta il 12 feb 2023

Vedi su GitHub
 (2 commenti) (3 reazioni) (0 assegnatari)Python (2500 fork)batch import
help wanted

Metriche repository

Star
 (27.526 star)
Metriche merge PR
 (Nessuna PR mergiata in 30 g)

Descrizione

It seems that I already used many tricks but training still OOM for 8GB gpu. But inference is good now.

This is strange because I know some textural inversion or dreambooth can be trained on 8GB.

What is the secrect of Automatic1111's optimization? Although xformers may help a bit, the currect sliced attention should require even smaller mem than xformers.

Does it make sence to move text encoder and vae outside gpu when training?

Guida contributor