Dave Berenbaum
Dave Berenbaum
> However in the "train stage" we use a _ML Framework_ + _ML Logger_ integration that `logs` additional information to the the tracking server. This is the information I referred...
> What are the examples of model object metadata that will need params files, not metrics? I'd be interested to hear what @daavoo thinks. In the past, I have logged...
> A) If `checkpoint: true` is set in DVC and `model_file` is None in DVCLive: we use the value of the output as `model_file` . What do you mean by...
It makes sense to me. It still seems slightly awkward to have to define the path in `dvc.yaml`. It would be great if this wasn't needed at all, and the...
Putting as p2 for now. Feel free to follow up if you disagree.
> During my evaluation stage, I plot a simple line plot where the x values are predefined, and the y values are calculated from truth and predicted values, and are...
Related: https://github.com/iterative/dvclive/pull/322#discussion_r990268395. Based on that discussion, it probably makes more sense to introduce a separate method here like `log_custom_plot` since the current one is so sklearn-focused. The arguments can mostly...
@sirily If I understand you correctly, your solution from the community gems then solves two different issues you have: 1. Unlike other frameworks, PyTorch Lightning logging does not expose hooks...
I think it's easy enough for users to add integrations as needed (or for the dvc team to add them in response to demand), so it's probably not worthwhile to...
See #25 for more discussion of dependency management.