experiment-management: remove warning about no parameter validation

[mattlbeck]

Per iterative/dvc#6521

Covers feature change iterative/dvc#6521. Parameters will now be validated.

Related issue iterative/dvc#5477

[mattlbeck]

This simply removes the warning about lack of parameter validation at the moment.

[dberenbaum]

Do you mind marking this as a draft until iterative/dvc#5477 is merged?

[jorgeorpinel]

This looks like it's all we need here actually so I'll make it a regular PR and just label it. Thanks @mattlbeck

Actually...

[jorgeorpinel]

So far so good but we need a few other changes:

• The text in Updating experiment parameters on-the-fly should say something like "-S option takes an existing parameter name and"...
• Optinally/Additionally, the sample command block under that same text could have dvc param list and it's output (showing model.learning_rate among 1-2 other params) before the exp run --set-param model.learning_rate command).
• Corresponding update to https://dvc.org/doc/command-reference/exp/run#options
• Same in https://dvc.org/doc/command-reference/exp/run#example-modify-parameters-on-the-fly

Thanks

[mattlbeck]

* Optinally/Additionally, the sample command block under that same text could have `dvc param list` and it's output (showing `model.learning_rate` among 1-2 other params) before the `exp run --set-param model.learning_rate` command).

@jorgeorpinel unless I am missing something dvc param list doesn't exist. As an alternative we could demonstrate with a cat params.yaml?

[jorgeorpinel]

we could demonstrate with a cat params.yaml?

Yes @mattlbeck sorry, I made up a command haha. I applied your idea. The last 2 check boxes in #2824 (review) are still pending.

[jorgeorpinel]

Ah no, I see everything is addressed. Thanks again @mattlbeck !

[jorgeorpinel]

I merged #2944 but this is still open 🤔 and now has more conflicts...

[jorgeorpinel]

merged #2944 but this is still open

I think Restyled is having some issues (again)... After merging it to master, this PR showed conflicts with the formatted bits (should've been closed by that). And now that I solved the conflicts (used master versions) it created another PR with 0 changes (how do you even do that?): #2945

[jorgeorpinel]

how do you even do that?

My master merge wasn't correct -- left an accidental trailing space in there. The Restyled PR changed it back to the already merged version in master. (That's how.)

Now I fixed the formatting here and there are no changes in this PR... 🤷 merging!