milesial
milesial
Oops, forgot about this PR and deleted my fork... I'll reopen. As far as I know we're still waiting for https://github.com/pytorch/pytorch/pull/97582 to reland for NVTX3 to be integrated, and then...
In the process of fixing tests I discovered and fixed a bug where the scheduler wouldn't match its optimizer when multiple optimizers are instantiated with frequencies. Now the optimizers and...
@carmocca Ready for final review
The way I fixed the `tests/tests_pytorch/models/test_hooks.py::test_trainer_model_hook_system_fit[True-kwargs1]` tests is very flaky, so I'd appreciate if someone more familiar with these tests comes up with a better fix. edit: seems like it...
I also modified `native_amp.py` in both lightning_fabric and pytorch_lightning. It doesn't seem like the lightning_fabric one is called for a typical workflow, so I don't know if it's the right...
Hi, it doesn't really matter if you use 1 or 2 classes for binary classification. 1 will use a single channel mask then sigmoid and threshold. 2 will use a...
Hi, can you try the latest master version? Custom dataset support was improved.
You should be able to follow the `n_classes == 1` code path where masks have one channel in [0, 1]
The checkpoints contain the model weights yes. You should use the checkpoint for the epoch you prefer (better accuracy), usually the latest.
Hi, can you try the latest master version? Custom dataset support was improved.