Project for developing knowledge maps via deep text embeddings.
To start a development server
pip install -r requirements.txt
gunicorn --bind 0.0.0.0:8000 wsgi:app
static
: contains core files (data, config and html) for appnotebooks
: contain preprocessor for extracting text embeddings using BERT
Note: Use any port as available.
- Create pre-trained map configs for UMAP and t-SNE (update app to new version)
- Add text bar to automagically match top-k units (in vector space) as per query