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.

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

Unable to change val dataloader batch size in qat training #1115

Closed
haritsahm opened this issue Jun 1, 2023 · 2 comments
Closed

Unable to change val dataloader batch size in qat training #1115

haritsahm opened this issue Jun 1, 2023 · 2 comments
Labels
🐛 Bug Something isn't working

Comments

@haritsahm
Copy link

I've been working all day trying to figure out how to export qat model for different batch size. So from what I understand, I just need to override dataset_params.val_dataloader_params.batch_size so that I can use different batch size to export the onnx model as defined in sg_trainer.py#L2250.

After a while, I know that it uses the same batch_size as the train_dataloader_params.batch_size after applying pre_launch_callback function because the val_dataloader_params uses/copy train_dataloader_params in pre_launch_callbacks.py#L397

Is there any particular reason why you override it with train_dataloader_params?

@spsancti spsancti added the 🐛 Bug Something isn't working label Jun 1, 2023
@spsancti
Copy link
Contributor

spsancti commented Jun 1, 2023

Looks like a copy/paste error to me
@shaydeci @Louis-Dupont do you know it it has some special meaning here?

@shaydeci
Copy link
Contributor

shaydeci commented Jun 5, 2023

Nope, typo. Fixed in #1129 and will be included in the upcoming release.

@shaydeci shaydeci closed this as completed Jun 5, 2023
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
🐛 Bug Something isn't working
Projects
None yet
Development

No branches or pull requests

3 participants