-
Notifications
You must be signed in to change notification settings - Fork 32
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
RuntimeError: shape '[1, 512, 1, 32, 2]' is invalid for input of size 16448 #21
Comments
Hi~
|
Hi @PeizeSun thanks, and apologies for not having provided more details. I've pre-computed the codes with (removing the ten_crop flag for debug):
The toy dataset is single class, available from https://www.joligen.com/datasets/butterflies.tar The generated codes in the
From printing the shapes, features are (correctly afaik) of shape My diff on the training code is below, I've downsized the training to a single GPU for debug purposes.
I run training with
The full error is below:
|
the same issue |
I first run the following command to generate codes on the imagenet dataset and then run the following command to train it raises an error in the
I have printed the size of variables before line if i comment out @PeizeSun could you help to solve this problem? thanks. |
I have met the same issue. |
Hi, thanks for the interesting work.
I'm playing a bit with the code on a simple single-class dataset of 256x256 images, and I've modified basic things (imagenet hardcoded numbers, etc...).
I'm hitting the error above on the rope embedding:
Went chasing the issue, and it seems this is due to a mismatch between the precomputed
freqs_cis
and the reshaping of the attention vectors. This mismatch appears to mostly be due to the number of augmentations (I went from 10 to 2 during debug).If this error rings a bell, I'd appreciate any hint :) I see how to fix it with a hack (reducing aug to none), but I believe something else is wrong, otherwise it wouldn't work at all.
Thanks!
The text was updated successfully, but these errors were encountered: