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

[tests] remove pt_tf equivalence tests (#36253)

Browse Source
This commit is contained in:
Joao Gante
2025-02-19 11:55:11 +00:00
committed by GitHub
parent 1a81d774b1
commit 0863eef248
60 changed files with 56 additions and 2438 deletions
Download Patch File Download Diff File Expand all files Collapse all files

16
tests/models/vit_mae/test_modeling_tf_vit_mae.py
View File

@@ -270,22 +270,6 @@ class TFViTMAEModelTest(TFModelTesterMixin, PipelineTesterMixin, unittest.TestCa
output_for_kw_input = model(**inputs_np, noise=noise)
self.assert_outputs_same(output_for_dict_input, output_for_kw_input)
# overwrite from common since TFViTMAEForPretraining has random masking, we need to fix the noise
# to generate masks during test
def check_pt_tf_models(self, tf_model, pt_model, tf_inputs_dict):
# make masks reproducible
np.random.seed(2)
num_patches = int((tf_model.config.image_size // tf_model.config.patch_size) ** 2)
noise = np.random.uniform(size=(self.model_tester.batch_size, num_patches))
tf_noise = tf.constant(noise)
# Add `noise` argument.
# PT inputs will be prepared in `super().check_pt_tf_models()` with this added `noise` argument
tf_inputs_dict["noise"] = tf_noise
super().check_pt_tf_models(tf_model, pt_model, tf_inputs_dict)
# overwrite from common since TFViTMAEForPretraining has random masking, we need to fix the noise
# to generate masks during test
def test_keras_save_load(self):

16
tests/models/vit_mae/test_modeling_vit_mae.py
View File

@@ -204,22 +204,6 @@ class ViTMAEModelTest(ModelTesterMixin, PipelineTesterMixin, unittest.TestCase):
config_and_inputs = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_pretraining(*config_and_inputs)
# overwrite from common since ViTMAEForPretraining has random masking, we need to fix the noise
# to generate masks during test
def check_pt_tf_models(self, tf_model, pt_model, pt_inputs_dict):
# make masks reproducible
np.random.seed(2)
num_patches = int((pt_model.config.image_size // pt_model.config.patch_size) ** 2)
noise = np.random.uniform(size=(self.model_tester.batch_size, num_patches))
pt_noise = torch.from_numpy(noise)
# Add `noise` argument.
# PT inputs will be prepared in `super().check_pt_tf_models()` with this added `noise` argument
pt_inputs_dict["noise"] = pt_noise
super().check_pt_tf_models(tf_model, pt_model, pt_inputs_dict)
def test_save_load(self):
config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()