GitLab

Whenever target_mapping is provided to the input, XLNet outputs two different attention streams.
Based on that the attention output would be on of the two:
- a list of tensors (usual case for most transformers)
- a list of 2-tuples of tensors, one tesor for each of attention streams
Docs and unit-tests have been updated

Mean does not exist in TF2

German DistilBERT

from_pretrained: convert DialoGPT format

Only init encoder_attention_mask if stack is decoder

Added Mish Activation Function

Fix CTRL past

Fix encode_plus

Add support for resumable downloads for HTTP protocol.

Consider do_lower_case in PreTrainedTokenizer