A simple repository that generates names using language modeling. This repo uses simple bigram model to complex sequential achitectures to predict the next character.
More and improved models are in progress. ⏳
- Bigram (one character predicts the next one with a lookup table of counts) (Loss: 2.4807)
- MLP, following Bengio et al. 2003 (Loss: 2.1681)
- GRU, following Kyunghyun Cho et al. 2014 (Loss: 2.0212)
- LSTM, following Graves et al. 2014 (Loss: 2.0164)
A heartfelt thanks to @karpathy