SUMIN/HuggingFace_transformer
1
0
Fork 0
Code Issues Pull Requests Actions 7 Packages Projects Releases Wiki Activity

Rename compute_loss in TF models (#15207)

Browse Source 
* Rename compute_loss to hf_compute_loss to avoid conflicts with the new Keras method

* make style

* Adding deprecation warning to `compute_loss`

* Fix sneaky reference to compute_loss

* Replace logger.warning with warnings.warn

* Clarifying warning and deprecation timeline
This commit is contained in:
Matt
2022-01-19 13:29:07 +00:00
committed by GitHub
parent d1f5ca1afd
commit 2708bfa127
35 changed files with 166 additions and 126 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
tests/test_modeling_tf_common.py
View File

@@ -1064,7 +1064,7 @@ class TFModelTesterMixin:
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
for model_class in self.all_model_classes:
model = model_class(config)
if getattr(model, "compute_loss", None):
if getattr(model, "hf_compute_loss", None):
# The number of elements in the loss should be the same as the number of elements in the label
prepared_for_class = self._prepare_for_class(inputs_dict.copy(), model_class, return_labels=True)
added_label = prepared_for_class[