Files
HuggingFace_transformer/transformers
Rémi Louf 3520be7824 create encoder attention mask from shape of hidden states
We currently create encoder attention masks (when they're not provided)
based on the shape of the inputs to the encoder. This is obviously
wrong; sequences can be of different lengths. We now create the encoder
attention mask based on the batch_size and sequence_length of the
encoder hidden states.
2019-12-09 11:19:45 +01:00
..
2019-12-04 18:40:52 -05:00
2019-11-27 11:07:22 -05:00
2019-12-07 07:46:14 -05:00
2019-11-26 13:08:12 -05:00
2019-10-09 11:07:43 +02:00
2019-11-05 13:31:58 -05:00
2019-11-29 11:25:37 -05:00
2019-10-29 17:10:20 +01:00
2019-11-26 13:08:12 -05:00
2019-11-29 11:25:37 -05:00
2019-11-11 10:15:14 -05:00
2019-11-11 10:15:14 -05:00