lllyasviel/ControlNet

Training still OOM on 8GB gpu.

Open

#21 ouverte le 12 févr. 2023

Voir sur GitHub
 (2 commentaires) (3 réactions) (0 assignés)Python (2 500 forks)batch import
help wanted

Métriques du dépôt

Stars
 (27 526 stars)
Métriques de merge PR
 (Aucune PR mergée en 30 j)

Description

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?

Guide contributeur