TD model performance report #235
Conversation
✅ Deploy Preview for silly-keller-664934 ready!
To edit notification comments on pull requests, go to your Netlify site settings. |
|
@ejm714 I'm pushing a first draft to test out the build process, make sure the paths to the images work, etc. If you want to have a first look at the content, please do. Or you can wait for a more polished version. |
Codecov Report
Additional details and impacted files@@ Coverage Diff @@
## master #235 +/- ##
========================================
- Coverage 87.2% 87.0% -0.2%
========================================
Files 28 28
Lines 1961 1962 +1
========================================
- Hits 1710 1707 -3
- Misses 251 255 +4
|
|
@glipstein The new page describing the performance of the time-dependent species model is at https://deploy-preview-235--silly-keller-664934.netlify.app/models/td-full-metrics/ This is probably not where it will live in the site structure; I just put it there for now so we can preview it. Please take a look and let me know what you think. @ejm714 Any thoughts about where this should go? |
I think it's in the right section, let's just move it below the Densepose entry and call it "African species performance" |
Co-authored-by: Emily Miller <[email protected]>
Co-authored-by: Emily Miller <[email protected]>
Co-authored-by: Emily Miller <[email protected]>
Co-authored-by: Emily Miller <[email protected]>
|
@glipstein I took your suggestion about reordering sections and made some revisions to make the text flow. Please take another look and merge if you think it's done. |
|
@ejm714 @glipstein One minor question: I put a hyphen in "time-distributed". Based on a quick search, the literature seems to be divided on this point. Any preference? |
|
No strong preference on the hyphen, think it looks good with it. I think this is looking really great @AllenDowney! The recall recall curve looks a little grainy to me -- maybe try exporting with a higher DPI? It might be useful to add a title as well so that's easier to skim the page and understand what the plot is showing. |
|
@ejm714 Added a title and generated higher resolution figures -- that looks better! |
|
|
||
| ## African forest model | ||
|
|
||
| The African species time-distributed model was trained using almost **250,000 videos from 14 countries** in West, Central, and East Africa. |
There was a problem hiding this comment.
the reason i had this as time_distributed was to mimic the models page
There was a problem hiding this comment.
I don't think a reader of this page (or the models page) would understand why it's in code font.
|
@AllenDowney don't worry about the tests, codecov is being flaky. once the last comments are addressed, we should be good to merge 🚀 |
Co-authored-by: glipstein <[email protected]>
…g/zamba into td-full-metric-docs
There was a problem hiding this comment.
@AllenDowney @ejm714 just confirming this isn't waiting on me to merge? Anything else you need me to do?
Excited to have this page!
|
@glipstein I was waiting for one last look and an approval. Now I will press the big green button. |
Closes https://github.com/drivendataorg/pjmf-zamba/issues/137