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

Remove a TF usage warning and rework the documentation #9756

Merged
merged 5 commits into from
Jan 27, 2021
Merged

Remove a TF usage warning and rework the documentation #9756

Changes from 1 commit
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
60c4618
Rework documentation
jplu Jan 22, 2021
467d43e
Update the template
jplu Jan 22, 2021
122b6be
Trigger CI
jplu Jan 22, 2021
a183844
Restore the warning but with the TF logger
jplu Jan 26, 2021
429fad8
Update convbert doc
jplu Jan 27, 2021
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
Prev Previous commit
Next Next commit
Restore the warning but with the TF logger
  • Loading branch information
@jplu
jplu committed Jan 27, 2021
commit a183844f07ee83e97d828f05276d2d2a802a9e19
15 changes: 15 additions & 0 deletions src/transformers/modeling_tf_utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -44,6 +44,7 @@


logger = logging.get_logger(__name__)
tf_logger = tf.get_logger()


class TFModelUtilsMixin:
Expand Down Expand Up @@ -278,9 +279,23 @@ def booleans_processing(config, **kwargs):

if "use_cache" in kwargs:
final_booleans["use_cache"] = kwargs["use_cache"] if kwargs["use_cache"] is not None else config.use_cache

else:
if (
kwargs["output_attentions"] is not None
or kwargs["output_hidden_states"] is not None
or ("use_cache" in kwargs and kwargs["use_cache"] is not None)
):
tf_logger.warn(
"The parameters `output_attentions`, `output_hidden_states` and `use_cache` cannot be updated when calling a model."
"They have to be set to True/False in the config object (i.e.: `config=XConfig.from_pretrained('name', output_attentions=True)`)."
)

final_booleans["output_attentions"] = config.output_attentions
final_booleans["output_hidden_states"] = config.output_hidden_states

if kwargs["return_dict"] is not None:
tf_logger.warn("The parameter `return_dict` cannot be set in graph mode and will always be set to `True`.")
final_booleans["return_dict"] = True

if "use_cache" in kwargs:
Expand Down