metrics: 1.0 update#1360
Conversation
This comment has been minimized.
This comment has been minimized.
jorgeorpinel
left a comment
jorgeorpinel
left a comment
There was a problem hiding this comment.
First review, mostly copy edits on index page. Same feedback applies on similar parts of the other 2 docs.
DVC 1.0 release, as well as other related cmd refs
jorgeorpinel
left a comment
jorgeorpinel
left a comment
There was a problem hiding this comment.
Approving so this can be merged but lots of the changes are mine at this point, some one else may want to do a separate review?
| DVC itself does not ascribe any specific meaning for these numbers. Usually | ||
| these numbers are produced by the model training or model evaluation code and | ||
| serve as a way to compare and pick the best performing experiment. | ||
| ### Types of metrics |
There was a problem hiding this comment.
this should go before description
There was a problem hiding this comment.
I didn't think you meant that literally earlier since it breaks our standard cmd ref structure. But OK, did it now in 3b898b3 (included also continuous and scalar terminology update discussed in Slack).
WDYT? I'm not so sure about this one...
There was a problem hiding this comment.
standard cmd ref structure.
we should not be following or trying to preserve it a cost of actual meaning and natural flow of the text. I don't know if this should go before or should be part of description, but introducing types and explaining the difference should be before explaining a particular type.
There was a problem hiding this comment.
Would be tricky putting it in the Description before the basic explanation unless we remove the Types of metrics header and just make it into the first paragraph.
But for now I think it's fine actually, since this change only affects the index.md of both metrics and plots, which is kind of a placeholder cmd ref anyway. Other indices are used to explain concepts more than as cmd refs also, like for cache and remote 🙂 so we can be more flexible about these in terms of doc structure going fwd.
| are produced by the model training or model evaluation code and serve as a way | ||
| to compare and pick the best performing experiment. | ||
|
|
||
| ### Default metric files |
There was a problem hiding this comment.
should become part of the description
There was a problem hiding this comment.
do we need a similar paragraph in plots?
There was a problem hiding this comment.
It is part of the description.
There was a problem hiding this comment.
do we need a similar paragraph in plots?
No, plots require specifying a target/datafile
There was a problem hiding this comment.
p.s. this led me to discover the following issues: treeverse/dvc/issues/3925 and treeverse/dvc/issues/3924
There was a problem hiding this comment.
it overlaps a lot:
For example:
Metric files can be added to
dvc.yamlwith the--metrics(-m) or
--metrics-no-cache(-M) options ofdvc run, or manually to themetrics
section of a stage indvc.yaml:
and
This kind of metrics can be defined with the
-m(--metrics) and-M
(--metrics-no-cache) options ofdvc run.
Why can't we make it a single paragraph?
anyway, probably not a big deal .. we can keep it like this for now
There was a problem hiding this comment.
do we need a similar paragraph in plots?
No, plots require specifying a target/datafile
Actually probably we'll need it but this will be done as a result of resolving treeverse/dvc/issues/3925 (I'll keep track).
it overlaps a lot:
Yeah I didn't notice before. Removed redundancy in 47ad311 (part of #1382)!
Why can't we make it a single paragraph?
I could but would probably want to completely remove the dvc.yaml example in that case and the cache:false note. Also not sure it makes sense to have this info before the Supported file formats section. No strong opinion, I'll do it if you think it's best. Please lmk
avoid terms "scalar" and "continuous"
5 participants
DVC 1.0 metrics