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

how to use TensorBoard? #7

Closed
win5923 opened this issue Oct 18, 2021 · 9 comments
Closed

how to use TensorBoard? #7

win5923 opened this issue Oct 18, 2021 · 9 comments

Comments

@win5923
Copy link

win5923 commented Oct 18, 2021

hello i don't know how to add scalar to TensorBoard? I want to do this kind of topic, hoping to improve some ExpRate, but I don’t know much about lightning TensorBoard.

@Green-Wood
Copy link
Owner

You can reference this line in lit_bttr.py file. Just pass your metric value to self.log function method.
For more details, you can see pytorch-lightning official logging documentation.

@win5923
Copy link
Author

win5923 commented Oct 18, 2021

thank you so much~

@win5923
Copy link
Author

win5923 commented Oct 21, 2021

image
Does this need to be added?

@Green-Wood
Copy link
Owner

You don't need to pass TensorboadLogger to Trainer, because this is the default behavior.

@win5923
Copy link
Author

win5923 commented Nov 17, 2021

image
hello again, i want to ask what are #[b,], # [b, 1, H, W] , # [b, H, W] , # [b, l] mean? thx

@Green-Wood
Copy link
Owner

Well, these are just some shape descriptions of these tensors.
For example:

b: batch size
1: we have just one input channel
H: height of input image
W: width of input image
l: length of latex tokens

@win5923
Copy link
Author

win5923 commented Nov 17, 2021

image

Excuse me, you already have batch above, why do you still need to define class batch below?

image

@Green-Wood
Copy link
Owner

Because different images may have different sizes, you need to collate images in one batch into a single tensor in Batch class.

@win5923
Copy link
Author

win5923 commented Nov 17, 2021

Thank you very much, I see.

@win5923 win5923 closed this as completed Feb 2, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants