Top pytorch-lightning issues

Predict method to label new data

Important discussion enhancement help wanted won't fix
Llannelongue picture Llannelongue·16 May 2020·6 Comments
👍3

`model.test()` can fail for `ddp` because `args` in `evaluation_forward` are malformed

Priority P0 bug / fix good first issue help wanted
mstewart141 picture mstewart141·9 May 2020·8 Comments

What is the current state of per iteration (not per epoch) schedulers support?

question
devforfu picture devforfu·7 May 2020·6 Comments

LightningModule does not subclass from torch.nn.Module which breaks depedency injection of models

bug / fix help wanted
jfelectron picture jfelectron·5 May 2020·4 Comments

Incompatible Trainer.test()

Priority P0 bug / fix help wanted
isvoboda picture isvoboda·3 May 2020·9 Comments

Model checkpoint claims test_step() not defined

bug / fix help wanted
MatthewMatero picture MatthewMatero·1 May 2020·5 Comments

Ditch Trainer percent arguments, make overfit great again

enhancement help wanted
iakremnev picture iakremnev·29 Apr 2020·3 Comments
👍11

Compatibilty with PyTorch Geometric

question
MarkusRosen picture MarkusRosen·22 Apr 2020·6 Comments

Bug and question about logging -- missing epoch, validation before train?

bug / fix help wanted
guydav picture guydav·17 Apr 2020·15 Comments

Test results not logged to tensorboard, since 0.7.3, this worked in 0.7.1

bug / fix help wanted
WSzP picture WSzP·10 Apr 2020·11 Comments
👀3
 Previous issues
Next issues