|
Skip loss backward and optimizer step if loss undefined for one batch
|
|
1
|
1491
|
February 22, 2021
|
|
Saving/loading LightningModule with injected network
|
|
1
|
2533
|
February 22, 2021
|
|
Loading models with huggingface Automodel.from_pretrained
|
|
1
|
1673
|
February 15, 2021
|
|
Get max steps inside configure_optimizers
|
|
2
|
4306
|
February 4, 2021
|
|
Global step loaded by "load_from_check_point" is wrong
|
|
1
|
1320
|
February 3, 2021
|
|
How to save model checkpoints every 1000 batches of data during training
|
|
2
|
3228
|
January 23, 2021
|
|
`.detach()` cannot stop backprop in `training_step`
|
|
4
|
2801
|
January 21, 2021
|
|
ValueError: optimizer got an empty parameter list
|
|
2
|
4062
|
January 13, 2021
|
|
OOM error due to tensor accumulation when trying to use functional metrics API
|
|
3
|
3358
|
January 12, 2021
|
|
Validation Error: The validation bar turns red and no callbacks are called, no checkpoints are saved
|
|
5
|
1478
|
December 21, 2020
|
|
EfficientNet SwishImplementation Error
|
|
2
|
1680
|
December 14, 2020
|
|
Mypy issues with lightning module
|
|
1
|
1701
|
December 8, 2020
|
|
TensorBoard Logger - no self.experiment?
|
|
1
|
2682
|
November 23, 2020
|
|
How to save hparams when not provided as argument (apparently assigning to hparams is not recomended)?
|
|
7
|
12796
|
November 5, 2020
|
|
Docker file is broken
|
|
1
|
513
|
November 3, 2020
|
|
BERT model throws error when used in Pytorch Lightning
|
|
2
|
3189
|
October 17, 2020
|
|
Loading a checkpoint for trainer.test
|
|
2
|
814
|
October 16, 2020
|
|
Hparams not restored when using load_from_checkpoint (default argument values are the problem?)
|
|
2
|
6739
|
October 8, 2020
|
|
How to use multiple train dataloaders with different lengths
|
|
1
|
7985
|
September 27, 2020
|
|
Combining loss for multiple dataloaders
|
|
9
|
4242
|
September 12, 2020
|
|
How to move new torch tensor to device automatically
|
|
1
|
2779
|
August 27, 2020
|