tensorflow / tensorflow/tensorboard
Feature request for ability to set reference lines
Nobody has claimed this yet.
- Dominant language
- TypeScript
- Stars
- 7.2k
- Forks
- 1.7k
- Avg merge
- 4d 22h
- Merged PRs (30d)
- 1
Description
It would be nice if there would be some simple way to add some sort of reference line to the scalar plots. E.g. for some particular tasks, it might be useful to consider the loss when predicting the mean signal or to have a reference to the threshold that corresponds to random predictions, human performance, state-of-the-art performance. Most of the references I can think of, could be represented by a single horizontal line, but maybe there could also be interest for more complex references.
For now, I just logged a fixed value using the tensorflow summaries, but it does look quite clumsy and it seems unnecessary to write this single value repeatedly in a long sequence just for the sake of having it in the plots. It would be nice to have some file in the log-directory that specifies references in some way and that can be added to specific plots (e.g. using a horizontal dotted line across the entire plot).
I might have overlooked something, but there seems to be no elegant way to incorporate this kind of references, and it would definitely be nice to have.
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start by locating the scalar plot rendering and TensorFlow summaries handling, then trace how data from the log-directory is incorporated into specific plots. The issue does not name files or tests, and the scope includes both fixed horizontal references and possible complex references. Done should include an agreed reference format and visible reference lines on the relevant scalar plots.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- typescript
- Domain
- data-visualization, frontend
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100