Word2Vec Demo

Training word embeddings in the browser using TensorFlow.js

Place your text below:

Specify model and parameters to generate dataset:

Window size
Tokens:
Training examples:
No. Token Freq
No. Input Output

View the input and output vectors to the model:

 

Embedding size
Learning rate
Epochs

The model be trained with categorical cross entropy loss function.

Specify parameters to run t-SNE:

Learning rate
Perplexity
Iterations

References:

Adapted from Andrej Karpathy's t-SNE CSV web demo

Theme from Bootstrap

Maintained by Raimi