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

Fix inverted conditional in TF common test! (#22540)

Browse Source 
* Fix inverted conditional in TF common test!

* Make the same change in the PT tests file

* Make sure hidden states for GPT2 have the same output shape in PT/TF

* Minor fix to PT implementation of token classification loss

* Skip loss equivalence test for TFHubert because it keeps overflowing to inf

* Compute LM loss for TF the (weird) way it's computed in PT

* Skip loss equivalence test for Wav2Vec2 for the same reason as Hubert

* Fix - don't try to access the hidden states property when output is a tuple
This commit is contained in:
Matt
2023-04-04 21:59:54 +01:00
committed by GitHub
parent 48fbd8fa2e
commit edb704b26e
7 changed files with 245 additions and 17 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
tests/test_modeling_common.py
View File

@@ -2030,7 +2030,7 @@ class ModelTesterMixin:
# For some models (e.g. base models), there is no label returned.
# Set the input dict to `None` to avoid check outputs twice for the same input dicts.
if set(pt_inputs_dict_with_labels.keys()).symmetric_difference(pt_inputs_dict.keys()):
if not set(pt_inputs_dict_with_labels.keys()).symmetric_difference(pt_inputs_dict.keys()):
pt_inputs_dict_with_labels = None
# Check we can load pt model in tf and vice-versa with model => model functions