Upgrade lucid to use tensorflow v2 APIs and make it work on Colab platform - #305
Upgrade lucid to use tensorflow v2 APIs and make it work on Colab platform#305ercaronte wants to merge 38 commits into
Conversation
Sorry, something went wrong.
…esting are passing. Tested on python 3.7 only.
All tox testing are passing. Tested on python 3.7 only.
All tox testing are passing for python 3.7 and 3.9 One exception with 3.9 is the test_channel_reducer_trivial that fails, but it seems a precision error.
All tox testing are passing for python 3.7 and 3.9 Fixed some small issues with some tests.
…ther minor fixes.
…ontent Adding missing images that are loaded by lucid notebook
…ivation-atlas-collect) that requires the imagenet dataset. Few other minor fixes.
Few other minor fixes in the lucid core python library.
…ontent Bugfix/missing notebook static content
…g off to later on the necessary fixes or refactoring.
Notes: 3D rendering is broken and pyopengl rendering code works only in colab (does not in kaggle).
… the lucid library.
…ate the sprite maps at this stage. Left as future work.
Upgrade lucid to use tensorflow v2 APIs and make it work on Colab platform
|
Check out this pull request on See visual diffs & provide feedback on Jupyter Notebooks. Powered by ReviewNB |
Sorry, something went wrong.
|
While this is being handled by the lucid team, I have merged these changes on my fork https://github.com/ercaronte/lucid. For anybody willing to use the lucid library on Colab and Kaggle feel free to use it by replacing the lucid package install with: !pip install --quiet git+https://github.com/ercaronte/lucid.git |
Sorry, something went wrong.
…ecated old functions that are no longer working.
|
A bit of an old PR but I noticed that the ReviewNB link is perma-loading. These notebook changes are reviewable in GitNotebooks: https://gitnotebooks.com/tensorflow/lucid/pull/305 |
Sorry, something went wrong.
This PR is to update the lucid repository so that it is possible to use it in Colab, now that it does not longer support the magic command %tensorflow_version 1.x.
I have updated the python library python code and most of the notebooks.
Since the lucid gs bucket is down, I have also fixed some of the links to static content, pointing temporarely to my lucid fork. If the PR is accepted, it will be possible to update those links to point to the official repo.
As a future work to maintain this repo: