Command-line option for layer-specific learning rates for non-sgd optimization - Githubissues

harvardnlp / seq2seq-attn

Sequence-to-sequence model with LSTM encoder/decoders and attention

http://nlp.seas.harvard.edu/code

MIT License

1.26k stars 278 forks source link

Command-line option for layer-specific learning rates for non-sgd optimization #82

Closed swiseman closed 7 years ago