Dave Berenbaum

Sorry, I missed this mention. No strong opinion on whether to have an index page. LGTM.

Looks really good overall! Mostly need to figure out the resume/checkpoints content and refine the dvc integration a little.

@jorgeorpinel What's the status of this PR? I see some unresolved discussions, but it's also approved, so up to you at this point I think.

> If VSCode allows users to provide (and perhaps persist the values) these custom targets, it would enable a quick way of exploring the Experiment Table and Plots views without...

The latest proposal on this is in https://www.notion.so/iterative/Python-Notebook-Experiments-5a92a6f1677841bfb3613983f75a60b2. It looks likely that we will be prioritizing top-level metrics and params for this reason.

Yes, although it's not that hard a requirement if there are other ideas for how to do it. If it is saved to `dvc.yaml`, it could be like a lightweight...

> Probably a controversial idea, but since we have the ribbon, how about we completely drop the toggling from experiments? This has my vote. What about going even further here...

Before creating the experiment ref, DVCLive should write to a `dvc.yaml` file inside the `Live` directory, which should configure parameters, metrics, and plots logged by DVCLive. This could happen either...

Do we need an issue in both? In dvc-remder to add something like `render_ipython` and here to add support in DVCLive?

Related: https://github.com/iterative/dvclive/issues/89. @shcheklein Are you encountering an issue when using both `tf.keras.callbacks.ModelCheckpoint` and `DvcLiveCallback`? Or you want to avoid needing `tf.keras.callbacks.ModelCheckpoint`?