Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Upgrade Roberta tokenizer #1821

Merged
merged 13 commits into from
Mar 28, 2022
Merged

Upgrade Roberta tokenizer #1821

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
13 commits
Select commit Hold shift + click to select a range
4b71d15
update roberta
yingyibiao Mar 21, 2022
f6bbcf2
update roberta tokenizer
yingyibiao Mar 23, 2022
501f674
update roberta tokenizer
yingyibiao Mar 23, 2022
95c7994
Merge branch 'develop' into roberta
yingyibiao Mar 23, 2022
4d3073f
Merge branch 'develop' into roberta
yingyibiao Mar 24, 2022
d48441c
update
yingyibiao Mar 24, 2022
b6c0d98
Merge branch 'roberta' of github.com:yingyibiao/PaddleNLP into roberta
yingyibiao Mar 24, 2022
cd9a79f
Merge branch 'develop' into roberta
yingyibiao Mar 25, 2022
b234b4a
Merge branch 'develop' into roberta
yingyibiao Mar 25, 2022
786b98b
Merge branch 'develop' into roberta
yingyibiao Mar 26, 2022
e53c146
update
yingyibiao Mar 27, 2022
d77d281
Merge branch 'develop' into roberta
yingyibiao Mar 28, 2022
986f0b8
Merge branch 'develop' into roberta
yingyibiao Mar 28, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 2 additions & 1 deletion paddlenlp/transformers/auto/tokenizer.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -48,7 +48,8 @@
("MBartTokenizer", "mbart"),
("MPNetTokenizer", "mpnet"),
("NeZhaTokenizer", "nezha"),
("RobertaTokenizer", "roberta"),
("RobertaChineseTokenizer", "roberta"),
("RobertaBPETokenizer", "roberta"),
("RoFormerTokenizer", "roformer"),
("ReformerTokenizer", "reformer"),
("SqueezeBertTokenizer", "squeezebert"),
Expand Down
4 changes: 3 additions & 1 deletion paddlenlp/transformers/bert/modeling.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -499,7 +499,9 @@ def forward(self,
else:
if attention_mask.ndim == 2:
# attention_mask [batch_size, sequence_length] -> [batch_size, 1, 1, sequence_length]
attention_mask = attention_mask.unsqueeze(axis=[1, 2])
attention_mask = attention_mask.unsqueeze(
axis=[1, 2]).astype(paddle.get_default_dtype())
attention_mask = (1.0 - attention_mask) * -1e4
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

这里和其他模型中ndim==2的行为是一致的不，是已经统一了mask的这个语义了吗

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

还没有统一，这个pr统一了 bert 和 roberta


embedding_output = self.embeddings(
input_ids=input_ids,
Expand Down
Loading