output
20062026
most citedTransformers are RNNs: Fast Autoregressive Transformers with Linear Attention

310 citations

Showing 2017 · cs.CLShow all

2 papers · 2 filters