ByteDance-Seed / ByteDance-Seed/Bagel
OOM while saving the finetune checkpoints
Open
- Dominant language
- Python
- Stars
- 6.2k
- Forks
- 545
- PR merge metrics
- No merged PRs in 30d
Description
I am doing the t2i finetune following the suggested arguments settings with 8xH800. It can successfully train. When saving the checkpoint, it encounters with OOM sometimes. I was wondering if there is some methods to reduce the VRAM usage, I checked the batch_size which is already set as 1.
Regards.
Contributor guide
No contributing guide indexed for this repository
Assessment
This issue has not been assessed yet.