Found 1 document(s)
Page 1 of 1
-
Recurrent transformers reuse weights to perform deep computation with fewer parameters. We study an LSTM-style transformer that also shares one key--value cache across its recurrent core. At width 1024, it uses 40.4% few...