Use random_attention_mask for TF tests (#16517)

* use random_attention_mask for TF tests * Fix for TFCLIP test (for now). Co-authored-by: ydshieh <[email protected]>
huggingface · Apr 1, 2022 · 2199382 · 2199382
1 parent 823dbf8
commit 2199382
Show file tree

Hide file tree

Showing 29 changed files with 62 additions and 57 deletions.
diff --git a/...plate-{{cookiecutter.modelname}}/test_modeling_tf_{{cookiecutter.lowercase_modelname}}.py b/...plate-{{cookiecutter.modelname}}/test_modeling_tf_{{cookiecutter.lowercase_modelname}}.py
@@ -21,7 +21,7 @@
 from transformers.testing_utils import require_tf, slow
 
 from ..test_configuration_common import ConfigTester
-from ..test_modeling_tf_common import TFModelTesterMixin, floats_tensor, ids_tensor
+from ..test_modeling_tf_common import TFModelTesterMixin, floats_tensor, ids_tensor, random_attention_mask
 
 
 if is_tf_available():
@@ -92,7 +92,7 @@ def prepare_config_and_inputs(self):
 
         input_mask = None
         if self.use_input_mask:
-            input_mask = ids_tensor([self.batch_size, self.seq_length], vocab_size=2)
+            input_mask = random_attention_mask([self.batch_size, self.seq_length])
 
         token_type_ids = None
         if self.use_token_type_ids:

diff --git a/tests/albert/test_modeling_tf_albert.py b/tests/albert/test_modeling_tf_albert.py
@@ -21,7 +21,7 @@
 from transformers.testing_utils import require_tf, slow
 
 from ..test_configuration_common import ConfigTester
-from ..test_modeling_tf_common import TFModelTesterMixin, ids_tensor
+from ..test_modeling_tf_common import TFModelTesterMixin, ids_tensor, random_attention_mask
 
 
 if is_tf_available():
@@ -96,7 +96,7 @@ def prepare_config_and_inputs(self):
 
         input_mask = None
         if self.use_input_mask:
-            input_mask = ids_tensor([self.batch_size, self.seq_length], vocab_size=2)
+            input_mask = random_attention_mask([self.batch_size, self.seq_length])
 
         token_type_ids = None
         if self.use_token_type_ids:

diff --git a/tests/bert/test_modeling_tf_bert.py b/tests/bert/test_modeling_tf_bert.py
@@ -21,7 +21,7 @@
 from transformers.testing_utils import require_tf, slow
 
 from ..test_configuration_common import ConfigTester
-from ..test_modeling_tf_common import TFModelTesterMixin, floats_tensor, ids_tensor
+from ..test_modeling_tf_common import TFModelTesterMixin, floats_tensor, ids_tensor, random_attention_mask
 from ..utils.test_modeling_tf_core import TFCoreModelTesterMixin
 
 
@@ -96,7 +96,7 @@ def prepare_config_and_inputs(self):
 
         input_mask = None
         if self.use_input_mask:
-            input_mask = ids_tensor([self.batch_size, self.seq_length], vocab_size=2)
+            input_mask = random_attention_mask([self.batch_size, self.seq_length])
 
         token_type_ids = None
         if self.use_token_type_ids:

diff --git a/tests/clip/test_modeling_tf_clip.py b/tests/clip/test_modeling_tf_clip.py
@@ -301,6 +301,12 @@ def prepare_config_and_inputs(self):
         input_mask = None
         if self.use_input_mask:
             input_mask = random_attention_mask([self.batch_size, self.seq_length])
+            # make sure the first token has attention mask `1` to ensure that, after combining the causal mask, there
+            # is still at least one token being attended to for each batch.
+            # TODO: Change `random_attention_mask` in PT/TF/Flax common test file, after a discussion with the team.
+            input_mask = tf.concat(
+                [tf.ones_like(input_mask[:, :1], dtype=input_mask.dtype), input_mask[:, 1:]], axis=-1
+            )
 
         config = self.get_config()
 

diff --git a/tests/convbert/test_modeling_tf_convbert.py b/tests/convbert/test_modeling_tf_convbert.py
@@ -20,7 +20,7 @@
 from transformers.testing_utils import require_tf, slow
 
 from ..test_configuration_common import ConfigTester
-from ..test_modeling_tf_common import TFModelTesterMixin, ids_tensor
+from ..test_modeling_tf_common import TFModelTesterMixin, ids_tensor, random_attention_mask
 
 
 if is_tf_available():
@@ -94,7 +94,7 @@ def prepare_config_and_inputs(self):
 
         input_mask = None
         if self.use_input_mask:
-            input_mask = ids_tensor([self.batch_size, self.seq_length], vocab_size=2)
+            input_mask = random_attention_mask([self.batch_size, self.seq_length])
 
         token_type_ids = None
         if self.use_token_type_ids:

diff --git a/tests/ctrl/test_modeling_tf_ctrl.py b/tests/ctrl/test_modeling_tf_ctrl.py
@@ -20,7 +20,7 @@
 from transformers.testing_utils import require_tf, slow
 
 from ..test_configuration_common import ConfigTester
-from ..test_modeling_tf_common import TFModelTesterMixin, ids_tensor
+from ..test_modeling_tf_common import TFModelTesterMixin, ids_tensor, random_attention_mask
 
 
 if is_tf_available():
@@ -69,7 +69,7 @@ def prepare_config_and_inputs(self):
 
         input_mask = None
         if self.use_input_mask:
-            input_mask = ids_tensor([self.batch_size, self.seq_length], vocab_size=2)
+            input_mask = random_attention_mask([self.batch_size, self.seq_length])
 
         token_type_ids = None
         if self.use_token_type_ids:

diff --git a/tests/deberta/test_modeling_tf_deberta.py b/tests/deberta/test_modeling_tf_deberta.py
@@ -20,7 +20,7 @@
 from transformers.testing_utils import require_tf, slow
 
 from ..test_configuration_common import ConfigTester
-from ..test_modeling_tf_common import TFModelTesterMixin, ids_tensor
+from ..test_modeling_tf_common import TFModelTesterMixin, ids_tensor, random_attention_mask
 
 
 if is_tf_available():
@@ -92,7 +92,7 @@ def prepare_config_and_inputs(self):
 
         input_mask = None
         if self.use_input_mask:
-            input_mask = ids_tensor([self.batch_size, self.seq_length], vocab_size=2)
+            input_mask = random_attention_mask([self.batch_size, self.seq_length])
 
         token_type_ids = None
         if self.use_token_type_ids:

diff --git a/tests/deberta_v2/test_modeling_tf_deberta_v2.py b/tests/deberta_v2/test_modeling_tf_deberta_v2.py
@@ -20,7 +20,7 @@
 from transformers.testing_utils import require_tf, slow
 
 from ..test_configuration_common import ConfigTester
-from ..test_modeling_tf_common import TFModelTesterMixin, ids_tensor
+from ..test_modeling_tf_common import TFModelTesterMixin, ids_tensor, random_attention_mask
 
 
 if is_tf_available():
@@ -95,7 +95,7 @@ def prepare_config_and_inputs(self):
 
         input_mask = None
         if self.use_input_mask:
-            input_mask = ids_tensor([self.batch_size, self.seq_length], vocab_size=2)
+            input_mask = random_attention_mask([self.batch_size, self.seq_length])
 
         token_type_ids = None
         if self.use_token_type_ids:

diff --git a/tests/distilbert/test_modeling_tf_distilbert.py b/tests/distilbert/test_modeling_tf_distilbert.py
@@ -20,7 +20,7 @@
 from transformers.testing_utils import require_tf, slow
 
 from ..test_configuration_common import ConfigTester
-from ..test_modeling_tf_common import TFModelTesterMixin, ids_tensor
+from ..test_modeling_tf_common import TFModelTesterMixin, ids_tensor, random_attention_mask
 
 
 if is_tf_available():
@@ -70,7 +70,7 @@ def prepare_config_and_inputs(self):
 
         input_mask = None
         if self.use_input_mask:
-            input_mask = ids_tensor([self.batch_size, self.seq_length], vocab_size=2)
+            input_mask = random_attention_mask([self.batch_size, self.seq_length])
 
         sequence_labels = None
         token_labels = None

diff --git a/tests/dpr/test_modeling_tf_dpr.py b/tests/dpr/test_modeling_tf_dpr.py
@@ -19,7 +19,7 @@
 from transformers.testing_utils import require_tf, slow
 
 from ..test_configuration_common import ConfigTester
-from ..test_modeling_tf_common import TFModelTesterMixin, ids_tensor
+from ..test_modeling_tf_common import TFModelTesterMixin, ids_tensor, random_attention_mask
 
 
 if is_tf_available():
@@ -94,9 +94,8 @@ def prepare_config_and_inputs(self):
 
         input_mask = None
         if self.use_input_mask:
-            input_mask = ids_tensor(
-                [self.batch_size, self.seq_length], vocab_size=2
-            )  # follow test_modeling_tf_ctrl.py
+            # follow test_modeling_tf_ctrl.py
+            input_mask = random_attention_mask([self.batch_size, self.seq_length])
 
         token_type_ids = None
         if self.use_token_type_ids:

diff --git a/tests/electra/test_modeling_tf_electra.py b/tests/electra/test_modeling_tf_electra.py
@@ -20,7 +20,7 @@
 from transformers.testing_utils import require_tf, slow
 
 from ..test_configuration_common import ConfigTester
-from ..test_modeling_tf_common import TFModelTesterMixin, floats_tensor, ids_tensor
+from ..test_modeling_tf_common import TFModelTesterMixin, floats_tensor, ids_tensor, random_attention_mask
 
 
 if is_tf_available():
@@ -71,7 +71,7 @@ def prepare_config_and_inputs(self):
 
         input_mask = None
         if self.use_input_mask:
-            input_mask = ids_tensor([self.batch_size, self.seq_length], vocab_size=2)
+            input_mask = random_attention_mask([self.batch_size, self.seq_length])
 
         token_type_ids = None
         if self.use_token_type_ids:

diff --git a/tests/flaubert/test_modeling_tf_flaubert.py b/tests/flaubert/test_modeling_tf_flaubert.py
@@ -19,7 +19,7 @@
 from transformers.testing_utils import require_sentencepiece, require_tf, require_tokenizers, slow
 
 from ..test_configuration_common import ConfigTester
-from ..test_modeling_tf_common import TFModelTesterMixin, ids_tensor
+from ..test_modeling_tf_common import TFModelTesterMixin, ids_tensor, random_attention_mask
 
 
 if is_tf_available():
@@ -75,7 +75,7 @@ def __init__(
 
     def prepare_config_and_inputs(self):
         input_ids = ids_tensor([self.batch_size, self.seq_length], self.vocab_size)
-        input_mask = ids_tensor([self.batch_size, self.seq_length], 2, dtype=tf.float32)
+        input_mask = random_attention_mask([self.batch_size, self.seq_length], dtype=tf.float32)
 
         input_lengths = None
         if self.use_input_lengths:

diff --git a/tests/funnel/test_modeling_tf_funnel.py b/tests/funnel/test_modeling_tf_funnel.py
@@ -20,7 +20,7 @@
 from transformers.testing_utils import require_tf
 
 from ..test_configuration_common import ConfigTester
-from ..test_modeling_tf_common import TFModelTesterMixin, ids_tensor
+from ..test_modeling_tf_common import TFModelTesterMixin, ids_tensor, random_attention_mask
 
 
 if is_tf_available():
@@ -111,7 +111,7 @@ def prepare_config_and_inputs(self):
 
         input_mask = None
         if self.use_input_mask:
-            input_mask = ids_tensor([self.batch_size, self.seq_length], vocab_size=2)
+            input_mask = random_attention_mask([self.batch_size, self.seq_length])
 
         token_type_ids = None
         if self.use_token_type_ids:

diff --git a/tests/gpt2/test_modeling_tf_gpt2.py b/tests/gpt2/test_modeling_tf_gpt2.py
@@ -19,7 +19,7 @@
 from transformers.testing_utils import require_tf, slow
 
 from ..test_configuration_common import ConfigTester
-from ..test_modeling_tf_common import TFModelTesterMixin, floats_tensor, ids_tensor
+from ..test_modeling_tf_common import TFModelTesterMixin, floats_tensor, ids_tensor, random_attention_mask
 from ..utils.test_modeling_tf_core import TFCoreModelTesterMixin
 
 
@@ -74,7 +74,7 @@ def prepare_config_and_inputs(self):
 
         input_mask = None
         if self.use_input_mask:
-            input_mask = ids_tensor([self.batch_size, self.seq_length], vocab_size=2)
+            input_mask = random_attention_mask([self.batch_size, self.seq_length])
 
         token_type_ids = None
         if self.use_token_type_ids:

diff --git a/tests/gptj/test_modeling_tf_gptj.py b/tests/gptj/test_modeling_tf_gptj.py
@@ -20,7 +20,7 @@
 from transformers.testing_utils import require_tf, slow, tooslow
 
 from ..test_configuration_common import ConfigTester
-from ..test_modeling_tf_common import TFModelTesterMixin, ids_tensor
+from ..test_modeling_tf_common import TFModelTesterMixin, ids_tensor, random_attention_mask
 from ..utils.test_modeling_tf_core import TFCoreModelTesterMixin
 
 
@@ -70,7 +70,7 @@ def prepare_config_and_inputs(self):
 
         input_mask = None
         if self.use_input_mask:
-            input_mask = ids_tensor([self.batch_size, self.seq_length], vocab_size=2)
+            input_mask = random_attention_mask([self.batch_size, self.seq_length])
 
         token_type_ids = None
         if self.use_token_type_ids:

diff --git a/tests/layoutlm/test_modeling_tf_layoutlm.py b/tests/layoutlm/test_modeling_tf_layoutlm.py
@@ -21,7 +21,7 @@
 from transformers.testing_utils import require_tf, slow
 
 from ..test_configuration_common import ConfigTester
-from ..test_modeling_tf_common import TFModelTesterMixin, ids_tensor
+from ..test_modeling_tf_common import TFModelTesterMixin, ids_tensor, random_attention_mask
 
 
 if is_tf_available():
@@ -107,7 +107,7 @@ def prepare_config_and_inputs(self):
 
         input_mask = None
         if self.use_input_mask:
-            input_mask = ids_tensor([self.batch_size, self.seq_length], vocab_size=2)
+            input_mask = random_attention_mask([self.batch_size, self.seq_length])
 
         token_type_ids = None
         if self.use_token_type_ids:

diff --git a/tests/longformer/test_modeling_tf_longformer.py b/tests/longformer/test_modeling_tf_longformer.py
@@ -20,7 +20,7 @@
 from transformers.testing_utils import require_sentencepiece, require_tf, require_tokenizers, slow
 
 from ..test_configuration_common import ConfigTester
-from ..test_modeling_tf_common import TFModelTesterMixin, ids_tensor
+from ..test_modeling_tf_common import TFModelTesterMixin, ids_tensor, random_attention_mask
 
 
 if is_tf_available():
@@ -79,7 +79,7 @@ def prepare_config_and_inputs(self):
 
         input_mask = None
         if self.use_input_mask:
-            input_mask = ids_tensor([self.batch_size, self.seq_length], vocab_size=2)
+            input_mask = random_attention_mask([self.batch_size, self.seq_length])
 
         token_type_ids = None
         if self.use_token_type_ids:

diff --git a/tests/lxmert/test_modeling_tf_lxmert.py b/tests/lxmert/test_modeling_tf_lxmert.py
@@ -23,7 +23,7 @@
 from transformers.testing_utils import require_tf, slow
 
 from ..test_configuration_common import ConfigTester
-from ..test_modeling_tf_common import TFModelTesterMixin, ids_tensor
+from ..test_modeling_tf_common import TFModelTesterMixin, ids_tensor, random_attention_mask
 
 
 if is_tf_available():
@@ -124,7 +124,7 @@ def prepare_config_and_inputs(self):
 
         input_mask = None
         if self.use_lang_mask:
-            input_mask = ids_tensor([self.batch_size, self.seq_length], vocab_size=2)
+            input_mask = random_attention_mask([self.batch_size, self.seq_length])
         token_type_ids = None
         if self.use_token_type_ids:
             token_type_ids = ids_tensor([self.batch_size, self.seq_length], self.type_vocab_size)

diff --git a/tests/mobilebert/test_modeling_tf_mobilebert.py b/tests/mobilebert/test_modeling_tf_mobilebert.py
@@ -20,7 +20,7 @@
 from transformers.testing_utils import require_tf, slow
 
 from ..test_configuration_common import ConfigTester
-from ..test_modeling_tf_common import TFModelTesterMixin, ids_tensor
+from ..test_modeling_tf_common import TFModelTesterMixin, ids_tensor, random_attention_mask
 
 
 if is_tf_available():
@@ -114,7 +114,7 @@ def prepare_config_and_inputs(self):
 
             input_mask = None
             if self.use_input_mask:
-                input_mask = ids_tensor([self.batch_size, self.seq_length], vocab_size=2)
+                input_mask = random_attention_mask([self.batch_size, self.seq_length])
 
             token_type_ids = None
             if self.use_token_type_ids:

diff --git a/tests/mpnet/test_modeling_tf_mpnet.py b/tests/mpnet/test_modeling_tf_mpnet.py
@@ -20,7 +20,7 @@
 from transformers.testing_utils import require_tf, slow
 
 from ..test_configuration_common import ConfigTester
-from ..test_modeling_tf_common import TFModelTesterMixin, ids_tensor
+from ..test_modeling_tf_common import TFModelTesterMixin, ids_tensor, random_attention_mask
 
 
 if is_tf_available():
@@ -90,7 +90,7 @@ def prepare_config_and_inputs(self):
 
         input_mask = None
         if self.use_input_mask:
-            input_mask = ids_tensor([self.batch_size, self.seq_length], vocab_size=2)
+            input_mask = random_attention_mask([self.batch_size, self.seq_length])
 
         sequence_labels = None
         token_labels = None

diff --git a/tests/openai/test_modeling_tf_openai.py b/tests/openai/test_modeling_tf_openai.py
@@ -20,7 +20,7 @@
 from transformers.testing_utils import require_tf, slow
 
 from ..test_configuration_common import ConfigTester
-from ..test_modeling_tf_common import TFModelTesterMixin, ids_tensor
+from ..test_modeling_tf_common import TFModelTesterMixin, ids_tensor, random_attention_mask
 
 
 if is_tf_available():
@@ -70,7 +70,7 @@ def prepare_config_and_inputs(self):
 
         input_mask = None
         if self.use_input_mask:
-            input_mask = ids_tensor([self.batch_size, self.seq_length], vocab_size=2)
+            input_mask = random_attention_mask([self.batch_size, self.seq_length])
 
         token_type_ids = None
         if self.use_token_type_ids:

diff --git a/tests/rembert/test_modeling_tf_rembert.py b/tests/rembert/test_modeling_tf_rembert.py
@@ -20,7 +20,7 @@
 from transformers.testing_utils import require_tf, slow
 
 from ..test_configuration_common import ConfigTester
-from ..test_modeling_tf_common import TFModelTesterMixin, floats_tensor, ids_tensor
+from ..test_modeling_tf_common import TFModelTesterMixin, floats_tensor, ids_tensor, random_attention_mask
 
 
 if is_tf_available():
@@ -95,7 +95,7 @@ def prepare_config_and_inputs(self):
 
         input_mask = None
         if self.use_input_mask:
-            input_mask = ids_tensor([self.batch_size, self.seq_length], vocab_size=2)
+            input_mask = random_attention_mask([self.batch_size, self.seq_length])
 
         token_type_ids = None
         if self.use_token_type_ids: