Top pytorch-lightning issues

Unable to save model using torch.save in pytorch lightning version 0.10

bug / fix help wanted
shrinath-suresh picture shrinath-suresh·13 Oct 2020·7 Comments

TPU error

TPU bug / fix help wanted
ishgirwan picture ishgirwan·10 Oct 2020·13 Comments

ModelCheckpoint filename unable to use metrics that contain a slash

Priority P1 bug / fix enhancement help wanted
its-dron picture its-dron·9 Oct 2020·20 Comments

Deprecate EvalModelTemplatein favor of BoringModel and another simple model that does actually learn

API / design enhancement good first issue help wanted tests / CI
Borda picture Borda·3 Oct 2020·9 Comments

Current batch loss and mean reduced loss

question
Ceceu picture Ceceu·1 Oct 2020·3 Comments

How to store test_step outputs to file?

question won't fix
Ceceu picture Ceceu·28 Sep 2020·9 Comments

Checkpoints based on validation_step or validation_epoch_end

question won't fix
johngrabner picture johngrabner·28 Sep 2020·4 Comments
👍1

How to keep some LightningModule's parameters on cpu when using CUDA devices for training

question
David-AJ picture David-AJ·28 Sep 2020·17 Comments

incorrect batch_sizes when Dataloader returns a dict with multiple tensors.

Priority P0 bug / fix
gerardsn picture gerardsn·26 Sep 2020·21 Comments

Not able to train lightning with tpu

Working as intended question
soumochatterjee picture soumochatterjee·10 Sep 2020·17 Comments
 Previous issues
Next issues