File size: 161 Bytes
5fa1a76
1
A tokenizer stores the vocabulary for each model and provide methods for encoding and decoding strings in a list of token embedding indices to be fed to a model.