Skip to content

Issues: libffcv/ffcv-imagenet

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

11 Open 7 Closed
11 Open 7 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Assignee
Filter by who’s assigned
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Issues list

Error in val_loop
#22 opened Apr 26, 2024 by AndreaBrg
numpy.core._exceptions._ArrayMemoryError: Unable to allocate 2.73 TiB for an array with shape (124994103978,) and data type [('sample_id', '<u8'), ('ptr', '<u8'), ('size', '<u8’)]
#21 opened Feb 27, 2024 by TheDarkKnight-21th
yaml file for imagenet writter hyperparameters?
#17 opened Feb 1, 2023 by realliyifei
Potential minor speedup: Gaussian blur is a separable 2d convolution
#16 opened Jul 25, 2022 by lebrice
1
batch_size=1 causes error when Squeeze() is in the "label" pipeline
#13 opened Jun 27, 2022 by botcs
1
1
ValueError(“total size of new array must be changed”)
#12 opened May 31, 2022 by cindy-cheng0214
Training extremely slow
#11 opened Apr 22, 2022 by netw0rkf10w
4
How to enable Multi-GPU training (1 model, multiple GPUs) under the server with limited memory?
#10 opened Apr 20, 2022 by fantasysee
3
DEFAULT_CROP_RATIO wrong?
#9 opened Apr 10, 2022 by ByungKwanLee
1
PR for resuming training from checkpoint and validate model loaded from disk
#6 opened Feb 17, 2022 by aniketrege
train_loss variable is None and val_loss variable in log not showing up
#5 opened Feb 11, 2022 by rraju1
5
ProTip! Add no:assignee to see everything that’s not assigned.