Add activation grids notebook

[mehdidc]

Issue #3, reproducing activation grids (https://github.com/tensorflow/lucid/blob/master/notebooks/building-blocks/ActivationGrid.ipynb)

It's possible to try it here: https://colab.research.google.com/drive/1pEe-KmXeDJcWQYLOHwcMubS69wVVCHLe#scrollTo=xidm-QrXvL2X

Here are the results so far with inceptionv1 and layer mixed4d:

• reproduced: https://imgur.com/twmizR4
• original: https://imgur.com/eaYwEWR

Some remarks and a question:

• I added channel_reducer as is from the original repo
• default transforms in transform.py produce a different size (due to random scaling) each time it is called, then resampling to 224 is done after that to have a fixed size. Is that the same in lucid? I need to debug a little bit in the original repo to be sure to answer that, but please let me know if you have the answer. The reason I ask is that in this specific notebook, the cells in the grid are much smaller than 224. I added an argument "fixed_image_size" to handle this specific case where we want a fixed image size (after resampling) which is not 224.
• Since all layers are computed and with this commit we can accept smaller images, this means that it's possible to have an exception on higher layers because the image size is not big enough, but it should be fine as long as the layer we are interested in is computed, I handled this exception

[greentfrapp]

@mehdidc Thanks so much for the PR! Sample results look good! I'm a little occupied with a paper submission at the moment, but I will definitely take a look at this over the weekend.

[greentfrapp]

@mehdidc sorry for the late reply!

Re. the issue on smaller images: for Lucid, I believe because Tensorflow 1.x allows you to just retrieve a specific layer via sess.run, so there is no issue as long as the layer you want is computed. But in our case, we run the entire model, so torch raises an error when you hit the later layers. I think I'm okay with raising an exception for now, so your modification is probably alright!

Two small changes:

• Could you format the quoted part in the notebook as a quote? i.e. using ">" in markdown
• Could you add some tests for channel_reducer? You can run the tests with

$ coverage run --source . --omit setup.py -m pytest
$ coverage report -m

Should be good once those are done, I'll take care of adding it to the main README.

Thanks again for the help!!!

[mehdidc]

Hi @greentfrapp thanks for the remarks, I added some tests + updated notebook quotes + other minor modifications

[greentfrapp]

Thanks @mehdidc for the help! I think this looks good to merge and I'll add the title card for the notebook soon!

[greentfrapp]

Closes #3