Skip to content

[QEff. Finetuning]: Adding tests for PP in HF trainer stack#817

Open
quic-swatia wants to merge 1 commit intoquic:ft_experimentalfrom
quic-swatia:pp-test
Open

[QEff. Finetuning]: Adding tests for PP in HF trainer stack#817
quic-swatia wants to merge 1 commit intoquic:ft_experimentalfrom
quic-swatia:pp-test

Conversation

@quic-swatia
Copy link
Contributor

No description provided.

Signed-off-by: Swati Allabadi <sallabad@qti.qualcomm.com>
Copy link
Contributor

@quic-akuruvil quic-akuruvil left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Please correct lint error

tensor_val = torch.tensor(value, dtype=torch.float32)
assert torch.isfinite(tensor_val), f"{label} is not finite: {value}"
assert value > lo, f"{label} = {value:.4f} ≤ {lo}; expected a positive loss."
assert value < hi, f"{label} = {value:.4f} ≥ {hi}; loss appears to have diverged or is unreasonably large."
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Against what are we comparing here?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants