vak icon indicating copy to clipboard operation
vak copied to clipboard

A neural network framework for researchers studying acoustic communication

Results 176 vak issues
Sort by recently updated
recently updated
newest added

e.g. `checkpoint-best.pt` this should probably be a constant somewhere This way any logic for loading the "best" checkpoint is consistent across all model families

**Describe the bug** None of the values that we log during runs through the CLI are recorded correctly in tensorboard log files **To Reproduce** Train any model e.g. as described...

BUG

We will eventually have a datasets module but what we really have right now are "pipes" in the sense of https://github.com/pytorch/data#what-are-datapipes. We should disambiguate the meaning of "dataset" (static set...

- [x] Fix the order in which a learning curve iterates through durations of the training split subsets - [x] Make fixes so CI runs (e.g., add `from __future__ import...

This came up in this related bug: https://github.com/vocalpy/vak/issues/697#issuecomment-1742131596 It's surprising that we specify the parameters for the post-processing transformation as separate options in the `predict` section of the config, but...

Right now a model definition requires that any `metrics` be `Callable` `class`. See this line in `vak.models.definition`: https://github.com/vocalpy/vak/blob/206c7344a39ea5974db4ce4a361250d171505cf2/src/vak/models/definition.py#L260 This can lead to cryptic errors when implementing a model--see this discussion...

ENH: enhancement

**Describe the bug** When running with a GPU that has Tensor Cores I am getting the following warning: > You are using a CUDA device ('NVIDIA GeForce RTX 4090') that...

BUG

I fixed the API docs so the doc build would work again in a2c6ea1ed2d6963b9e066392454a06b5119349d7, but there are still a bunch of warnings when we run `sphinx-build`, and a glance at...

DOC: documentation

#693 fixes unit tests after the scipy paper, along with some bugs, and does some initial clean up on how we generate test data as well. But I see more...

This is part of the fix for #687. I think we should just be explicit we're using Lightning - [ ] Change 'device' option to 'accelarator' in config files -...