Learning tfjs-tsne
March 4, 2023 ยท View on GitHub
Just one of the things I'm learning. https://github.com/hchiam/learning
Get t-SNE output coordinates with tfjs-tsne and visualization powered by tfjs-vis.
Great explanation of SNE vs t-SNE (vs UMAP - which sounds better than t-SNE): https://towardsdatascience.com/visualizing-your-embeddings-4c79332581a9 - basically UMAP is better than t-SNE is better than SNE (which is related to the order they were invented).
Integration note: because tfjs-tsne and tfjs-vis are unmaintained and rely on an old version of tfjs that conflicts with the version of tfjs that models like universal-sentence-encoder rely on, I recommend swapping tfjs-tsne with umap-js and swapping tfjs-vis with chart.js so they all play nice together. Apparently UMAP is better than t-SNE anyways. My other repo that uses umap-js can be found here: https://github.com/hchiam/learning-tfjs-umap
Learn more about TensorFlow at https://github.com/hchiam/learning-tensorflow
Starting by testing out this repo
Run yarn global add parcel or npm install -g parcel, and then:
Using yarn: (triple-click to select all)
git clone https://github.com/hchiam/learning-tfjs-tsne.git && cd learning-tfjs-tsne && yarn;
yarn dev;
Or with npm: (triple-click to select all)
git clone https://github.com/hchiam/learning-tfjs-tsne.git && cd learning-tfjs-tsne && npm install;
npm run dev;
notes to self
First run yarn dev to create the /dist folder, and then run this:
yarn deploy