lucid
lucid copied to clipboard
A collection of infrastructure and tools for research in neural network interpretability.
Unfortunately, this requires a little refactoring.
Channels
Allow image params to use an arbitrary amount of channels, then ensure that image_shape uses that correctly.
User says: > I don't know if you ever work with non-square images, but I noticed height and width seem to be the wrong way round in param.image (and possibly...
For now, I'm just copying the code from this [notebook](https://colab.research.google.com/github/tensorflow/lucid/blob/master/notebooks/building-blocks/SemanticDictionary.ipynb#scrollTo=MEWC-UKdqRGC). I get an error when generating the visualization. python: 3.6.8 lucid: 0.3.8 svelte-cli:2.2.0 svelte:1.64.1 ``` --------------------------------------------------------------------------- TypeError Traceback (most recent...
Hi, maybe this is a silly question out of my ignorance, if it is pls pardon me. I cant save modelzoo .pb graphs like i do with keras models loaded...
Due to an underlying bug in PIL (python-pillow/Pillow#1475), our present save implementation can't save very large images eg. (28800, 28800, 3) image. We may want to look into an alternate...
Hi, I'm trying to import Pixelcnn++ (https://github.com/openai/pixel-cnn) into Lucid, specifically the pretrained cifar10 version. I've tried using both the current API and the experimental one documented in #152 - in...
I'm trying to open an autoencoder model I've trained myself, on Lucid, and I'm using as reference the notebook _[Importing a graph into modelzoo](https://colab.research.google.com/drive/1PPzeZi5sBN2YRlBmKsdvZPbfYtZI-pHl#scrollTo=0xH9C5m166BN)_. I'm mostly in doubt on how...
I was trying to import a custom model to use with lucid, and using the `tf.saved_model` loader. Is there support for importing a graph_def in this way, or am I...
I was wondering about the viability of only sampling the content/output texture and updating where it's necessary from the view frame for each iteration. More specifically to sample the content...