Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Reduce file size of LoRA models #397

Open
isaac091 opened this issue May 31, 2024 · 0 comments · May be fixed by #383
Open

Reduce file size of LoRA models #397

isaac091 opened this issue May 31, 2024 · 0 comments · May be fixed by #383
Assignees
@isaac091
Labels
optimization Model training/inferencing optimization

Comments

@isaac091
Copy link
Collaborator

Currently, because we always resize and train the embeddings of LoRA models, the main model file is 6.1 GB even though it doesn't need to store any more information than a normal model, which is 2.5 GB.
@isaac091 isaac091 self-assigned this May 31, 2024
@ddaspit ddaspit added the enhancement New feature or request label Jun 7, 2024
@isaac091 isaac091 added optimization Model training/inferencing optimization and removed enhancement New feature or request labels Jun 17, 2024
@isaac091 isaac091 linked a pull request Jun 19, 2024 that will close this issue
Save LoRA models as full merged models rather than adapters #383
Open
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@isaac091 isaac091

Labels
optimization Model training/inferencing optimization
Projects
SIL-NLP Research
Status: 👀 In review
Milestone
No milestone
Development

Successfully merging a pull request may close this issue.

Save LoRA models as full merged models rather than adapters sillsdev/silnlp
2 participants
@ddaspit @isaac091