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

Why divide the loss by the epoch index? #15

Open

2jungeuni opened this issue Oct 5, 2023 · 0 comments

2jungeuni commented Oct 5, 2023 •

edited

Loading

First, thank you for releasing very well organized code.

If you divide the loss by the epoch index, the loss will inevitably decrease as learning progresses.

Can you briefly explain the followings?

"self.metrics["train/loss"] += loss / (i + 1)" and "self.metrics["train/e_loss"] += e_loss / (i + 1)" in src/trainers/full.py
"self.metrics["val/loss"] += loss / (i + 1)" and "self.metrics["val/e_loss"] += e_loss / (i + 1)" in src/trainers/full.py

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