Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Add projector plugin colab #3423

Merged
merged 10 commits into from Apr 20, 2020
Merged

Add projector plugin colab #3423

merged 10 commits into from Apr 20, 2020

Conversation

hfiller
Copy link
Contributor

@hfiller hfiller commented Mar 24, 2020

Adds a colab demo demonstrating how to use the projector plugin locally.

Adds a colab demo demonstrating how to use the projector plugin locally.
@review-notebook-app
Copy link

Check out this pull request on  ReviewNB

You'll be able to see Jupyter notebook diff and discuss changes. Powered by ReviewNB.

@hfiller hfiller requested a review from GalOshri March 24, 2020 16:31
@hfiller hfiller marked this pull request as ready for review March 24, 2020 16:33
@review-notebook-app
Copy link

review-notebook-app bot commented Mar 26, 2020

View / edit / reply to this conversation on ReviewNB

GalOshri commented on 2020-03-26T02:35:18Z
----------------------------------------------------------------

Should be 2020.


hfiller commented on 2020-03-31T21:19:13Z
----------------------------------------------------------------

Done

@review-notebook-app
Copy link

View / edit / reply to this conversation on ReviewNB

GalOshri commented on 2020-03-26T02:35:19Z
----------------------------------------------------------------

The table here isn't rendering correctly.


@review-notebook-app
Copy link

review-notebook-app bot commented Mar 26, 2020

View / edit / reply to this conversation on ReviewNB

GalOshri commented on 2020-03-26T02:35:20Z
----------------------------------------------------------------

I think "highly dimensional embeddings" -> "high dimensional embeddings".

Also, "visuallizing" typo

An image here might be nice :).


@review-notebook-app
Copy link

review-notebook-app bot commented Mar 26, 2020

View / edit / reply to this conversation on ReviewNB

GalOshri commented on 2020-03-26T02:35:20Z
----------------------------------------------------------------

I don't think we need to say that we are using TF 2.

'tensorboard' -> 'TensorBoard' (also there is an extra '`')

It seems like we are not actually using the MNIST dataset below.


@review-notebook-app
Copy link

review-notebook-app bot commented Mar 26, 2020

View / edit / reply to this conversation on ReviewNB

GalOshri commented on 2020-03-26T02:35:21Z
----------------------------------------------------------------

Maybe start with "We will use a dataset of ..." just to make it read a bit more smoothly.


@review-notebook-app
Copy link

review-notebook-app bot commented Mar 26, 2020

View / edit / reply to this conversation on ReviewNB

GalOshri commented on 2020-03-26T02:35:22Z
----------------------------------------------------------------

Should we refer to it as just "embedding" or "embedding layer"? I think the layer is what provides the embeddings.

I think instead of "vector" we should explicitly say "embedding"

It might be worth linking to additional resources so people can learn about embeddings. E.g. https://www.tensorflow.org/tutorials/text/word_embeddings?hl=en


@review-notebook-app
Copy link

review-notebook-app bot commented Mar 26, 2020

View / edit / reply to this conversation on ReviewNB

GalOshri commented on 2020-03-26T02:35:22Z
----------------------------------------------------------------

"The layer we are embedding": This makes it seem like embedding is a verb we are applying to the layer. I think instead it is "The embedding layer".


@review-notebook-app
Copy link

review-notebook-app bot commented Mar 26, 2020

View / edit / reply to this conversation on ReviewNB

GalOshri commented on 2020-03-26T02:35:24Z
----------------------------------------------------------------

Can this be done by the Keras TensorBoard callback?

Is this the only way of saving embeddings?


hfiller commented on 2020-03-30T08:07:13Z
----------------------------------------------------------------

Currently yes, I believe the Keras TensorBoard callback is broken, I'm currently working on fixing this.

@review-notebook-app
Copy link

review-notebook-app bot commented Mar 26, 2020

View / edit / reply to this conversation on ReviewNB

GalOshri commented on 2020-03-26T02:35:24Z
----------------------------------------------------------------

We should add an image here otherwise the tutorial page won't have it (see how it is done as an example at https://github.com/tensorflow/tensorboard/blob/90227dcf7e7f900ba77b5b7ced1b14378a7ec98f/docs/hyperparameter_tuning_with_hparams.ipynb

I think it would be useful to add some explanation of why this is interesting/valuable. What does the user get from doing this? Is there some insight?


hfiller commented on 2020-03-31T21:20:30Z
----------------------------------------------------------------

Added some analysis including a comparison between Alfred Hitchcock and Wes Anderson. Let me know what you think

@GalOshri
Copy link
Contributor

Thank you for adding this! This tutorial will be super valuable for helping people learn about and get started with the projector plugin.

Can you also add it to https://github.com/tensorflow/tensorboard/blob/master/docs/_book.yaml? This enables it to show up in the left pane on the docs website. Let's put it under the hyperparameter tuning tutorial.

@@ -0,0 +1,337 @@
{
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

"Tensorboard" -> "TensorBoard"

"from your tensorflow projects from the logs in the specified directory log_dir " is a bit confusing. Perhaps something like "from the logs in the specifiedlog_dir directory "?


Reply via ReviewNB

@@ -0,0 +1,337 @@
{
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It might be worth including a sentence on how you identified the more closely associated words (search for a particular word, the neighbors are highlighted. These are words that are close to it in the embedding space).


Reply via ReviewNB

Copy link
Contributor

@GalOshri GalOshri left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

A few other comments left through ReviewNB but looks great once those are resolved!

Thank you for taking the time to do this and also for finding a great example.

docs/_book.yaml Outdated
@@ -23,6 +23,8 @@ upper_tabs:
path: /tensorboard/graphs
- title: "Hyperparameter tuning"
path: /tensorboard/hyperparameter_tuning_with_hparams
- title: "Projector plugin"
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We want to avoid using "plugin" in user documentation. How about "Embedding projector" or "Projector dashboard"?

@hfiller hfiller merged commit 57cc9dc into master Apr 20, 2020
caisq pushed a commit to caisq/tensorboard that referenced this pull request May 19, 2020
* Add projector plugin colab

* Adds a colab demo demonstrating how to use the projector plugin locally.

* Add embedding_projector image for colab

* Add colab to _book.yaml

* Add analysis to projector plugin

* Adding images for embedding projector colab

* Rename Projector plugin to Embedding projector
@hfiller hfiller deleted the hfiller-add-colab branch May 19, 2020 20:00
caisq pushed a commit that referenced this pull request May 27, 2020
* Add projector plugin colab

* Adds a colab demo demonstrating how to use the projector plugin locally.

* Add embedding_projector image for colab

* Add colab to _book.yaml

* Add analysis to projector plugin

* Adding images for embedding projector colab

* Rename Projector plugin to Embedding projector
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet

4 participants