tensorflow / tensorflow/tensorboard

Plugin design: Do inference to visualize network activations

Open
#527 7 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

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

Description

Is it conceivable that one could write a plugin that visualizes activations similar to the tool from Yosinski? I wonder, because it seems that one should rather write summaries, that are then processed by the backend, but if one wants to interactively choose inputs, that does not seem appropriate. Would it make sense to just configure the backend with the graph_def and a path to a checkpoint and then serve inferred activations? I am asking, because I currently work on a tool simliar to Yosinski's and integration into Tensorboard would be nice, but I am not sure if this complies with the intended design.

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.

Research direction

Start with the TensorBoard plugin example and the Yosinski deep-visualization-toolbox linked in the issue, then review the summaries and backend path described here. Establish whether interactive input selection can be supported with a graph_def and checkpoint, and document the resulting plugin integration direction.

Written by the indexing model from the issue text.

Assessment

Domain
backend, data-visualization, machine-learning
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
20/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.