lucid icon indicating copy to clipboard operation
lucid copied to clipboard

A collection of infrastructure and tools for research in neural network interpretability.

Results 85 lucid issues
Sort by recently updated
recently updated
newest added

_inspired by #101_ Consider printing a warning if final image input doesn't match the model's declared `input_shape`. More flexible architectures should still work, so we shouldn't consider that an error—but...

enhancement

🔬 *This is an experiment in doing radically open research. I plan to post all my work on this openly as I do it, tracking it in this issue. I'd...

research

Does lucid support to make my own TensorFlow model visualizations now?And how to make my own channel spritemaps? Thanks!

question
discussion

_inspired by #101_ Once we merge in new robustness transforms, `crop_to_size` should probably be in the default transforms. We still want to allow not passing it, but should probably warn...

enhancement

Likely along the lines of `load(url, as_image=True)` or similar API. For bonus points, try using mime-type first when the URL response comes with one? via @ncammarata

enhancement

* I'd really like to have a convenient way to evaluate a model once without going through the whole rigamarole of setting up a graph. * Maybe `model.isolated_eval(layer, input)`. *...

enhancement
lucid.modelzoo

🔬 *This is an experiment in doing radically open research. I plan to post all my work on this openly as I do it, tracking it in this issue. I'd...

research

🔬 *This is an experiment in doing radically open research. I plan to post all my work on this openly as I do it, tracking it in this issue. I'd...

research

- [ ] (Maybe) Run linter first - [ ] Run unit tests on both platforms - [ ] if tagged commit, run integration tests - [ ] if success,...

enhancement
infrastructure