ANN-Fundamentals
ANN-Fundamentals copied to clipboard
dlab-berkeley
1. Build ANN Classifier: We show how to build an ANN using Keras Sequential API, but make no mention of the Functional API for more customized architectures. Might be worthwhile...
1. Introduction: "Handwritten Digits Data Sets" link is broken. Needs to be updated. 2. Training and Testing the ANN classifier": Maybe add some best practice hyperparameters (e.g., Adam solver, ReLU...
1. First cell: As in Notebook 1, instead of using np.random.random to sample and then manually adjusting the range, we could use np.random.uniform instead ==> "np.random.uniform(low=-1, high=1, size=(4,4))", "np.random.uniform(low=-1, high=1,...
1. General: More of a subjective opinion, but maybe reduce the time/number of slides spent on background/history of ANNs. As a workshop, I would personally hope to see more applied...
1. Cell 1: The output of the weights/targets from the first cell is hard to read. Maybe a table that shows a direct comparison of how the model learns for...
When testing notebook 4 on datahub, using the full mnist dataset will run out of memory if using 1GB of RAM allocated for standard berkeley datahub accounts. I have already...
This repo needs links in the [README.md](https://github.com/dlab-berkeley/ANN-Fundamentals/blob/main/README.md) for being able to run this in Datahub and Binder using the [requirements.txt](https://github.com/dlab-berkeley/ANN-Fundamentals/blob/main/requirements.txt) file already included— which also needs version upgrades for security...
@pssachdeva @seanmperez the [requirements.txt](https://github.com/dlab-berkeley/ANN-Fundamentals/blob/main/requirements.txt) file is used with Binder and Datahub. The specific version dependencies should be updated and all the notebooks tested thoroughly afterwards by several people. https://github.com/dlab-berkeley/ANN-Fundamentals/security/dependabot ...
Instructions for installing on a laptop need to be added to the [README.md](https://github.com/dlab-berkeley/ANN-Fundamentals/blob/main/README.md) @pssachdeva @seanmperez
