tensorflow / tensorflow/tensorboard

Ability to construct description and/or add free text description for each RUN

Open
#4,661 1 comment 0 reactions 1 assignee View on GitHub

@nfelt is already working on this.

Since Feb 8, 2021.

theme:usability type:feature
Dominant language
TypeScript
Stars
7.2k
Forks
1.7k
Avg merge
4d 22h
Merged PRs (30d)
1

Description

Although user has the flexibility to construct log_dir to be informative for future reference, let's think about enhancing its usability by allowing user to easily define that for each run.

Let's assume that I'm running two trials with 2 optimizers adam & sgd. I wish to have the ability either to define my run based on some argument values dynamically including free text.

Assuming that such definition will be under tf.keras.callbacks.TensorBoard(), then this what I hope to have:

Sample for Enhancement
idea is to pass some arguments into the callback and TensorBoard save their values as run name. Example:

tf.keras.callbacks.TensorBoard(..., run_description=[epochs, date, time, 'while using',optimizer,], ...)

Based on above, user will get run name as: "100_20210211_1459_while using_adam" as an example while doing another run later using sgd, he will get run name as "100_20210211_15029_while using_sgd"

As the proposed run_description could accept free text also, then user may pass
tf.keras.callbacks.TensorBoard(..., run_description=['I am testing something'], ...)

Thanks for your efforts and hope to have it or better soon.

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.