edu
edu copied to clipboard
wandb
Educational materials on deep learning by Weights & Biases
!pip install git+https://github.com/PyTorchLightning/pytorch-lightning so install using github gets no error in collab
See this issue: https://github.com/huggingface/datasets/issues/2646 We should switch to using Artifacts for greater reliability -- but that may require tinkering with the `load_dataset` process.
the conv channel sizes are "splatted" in the config -- logged as `conv.1.size` -- this would be good to add to the models. i think a callback would be good...
The switch to normalizing the input data with batchnorm means we don't need students to edit the `DataModule` directly, so we can pull that code out -- it's distracting and...
if we inherit from `lu`, could add in more code there to check that things are working and provide more useful error messages also, the final output size should be...
it's better to have the students directly work with the PyTorch layers -- can later introduce the FC/Conv layers for the contest notebook
Right now, doesn't appropriately test whether there are values above 1 or below 0 because the examples don't add up to 1, which is what most folks test for. `np.array([-1,...
Would like to drop reliance on the W&B Hub, and Colab seems like the least-bad choice. ### Options: 1) **Local**. No lol. Installation issues are a dealbreaker. 2) **Binder**. Ephemerality...
- better naming and saving - add config metadata
this way, folks are always logged in before they get to the meat of the code