This repository contains a implementation of Continuous Bag of Words (CBOW) Word2Vec embedding model from scratch using Python numpy and jax libraries.
The dataset we used to train the model is "Alice’s Adventures in Wonderland" from Kaggle, which includes 3,332 sentences and 2,690 unique words. The dataset is stored in the data/alice folder and you can use it to train your model. Alternatively, you can also put your own dataset under data folder to train your own model (you may need slight code modification on the data file path).