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

Automatically create/update tiny models (#22275)

Browse Source 
* Automatically create or update tiny models

* Skip failed tests

* update workflow file

* use revision

---------

Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>
This commit is contained in:
Yih-Dar
2023-03-23 19:14:17 +01:00
committed by GitHub
parent a92e0ad2e2
commit e8cc02555e
12 changed files with 5025 additions and 1671 deletions
Download Patch File Download Diff File Expand all files Collapse all files

9
tests/models/altclip/test_modeling_altclip.py
View File

@@ -402,6 +402,15 @@ class AltCLIPModelTest(ModelTesterMixin, PipelineTesterMixin, unittest.TestCase)
test_resize_embeddings = False
test_attention_outputs = False
# TODO: Fix the failed tests when this model gets more usage
def is_pipeline_test_to_skip(
self, pipeline_test_casse_name, config_class, model_architecture, tokenizer_name, processor_name
):
if pipeline_test_casse_name == "FeatureExtractionPipelineTests":
return True
return False
def setUp(self):
self.model_tester = AltCLIPModelTester(self)

9
tests/models/audio_spectrogram_transformer/test_modeling_audio_spectrogram_transformer.py
View File

@@ -165,6 +165,15 @@ class ASTModelTest(ModelTesterMixin, PipelineTesterMixin, unittest.TestCase):
test_resize_embeddings = False
test_head_masking = False
# TODO: Fix the failed tests when this model gets more usage
def is_pipeline_test_to_skip(
self, pipeline_test_casse_name, config_class, model_architecture, tokenizer_name, processor_name
):
if pipeline_test_casse_name == "AudioClassificationPipelineTests":
return True
return False
def setUp(self):
self.model_tester = ASTModelTester(self)
self.config_tester = ConfigTester(self, config_class=ASTConfig, has_text_modality=False, hidden_size=37)

9
tests/models/blenderbot_small/test_modeling_blenderbot_small.py
View File

@@ -237,6 +237,15 @@ class BlenderbotSmallModelTest(ModelTesterMixin, GenerationTesterMixin, Pipeline
test_pruning = False
test_missing_keys = False
# TODO: Fix the failed tests when this model gets more usage
def is_pipeline_test_to_skip(
self, pipeline_test_casse_name, config_class, model_architecture, tokenizer_name, processor_name
):
if pipeline_test_casse_name == "TextGenerationPipelineTests":
return True
return False
def setUp(self):
self.model_tester = BlenderbotSmallModelTester(self)
self.config_tester = ConfigTester(self, config_class=BlenderbotSmallConfig)

9
tests/models/deta/test_modeling_deta.py
View File

@@ -183,6 +183,15 @@ class DetaModelTest(ModelTesterMixin, GenerationTesterMixin, PipelineTesterMixin
test_head_masking = False
test_missing_keys = False
# TODO: Fix the failed tests when this model gets more usage
def is_pipeline_test_to_skip(
self, pipeline_test_casse_name, config_class, model_architecture, tokenizer_name, processor_name
):
if pipeline_test_casse_name == "ObjectDetectionPipelineTests":
return True
return False
# special case for head models
def _prepare_for_class(self, inputs_dict, model_class, return_labels=False):
inputs_dict = super()._prepare_for_class(inputs_dict, model_class, return_labels=return_labels)

9
tests/models/ernie_m/test_modeling_ernie_m.py
View File

@@ -250,6 +250,15 @@ class ErnieMModelTest(ModelTesterMixin, PipelineTesterMixin, unittest.TestCase):
)
test_torchscript = False
# TODO: Fix the failed tests when this model gets more usage
def is_pipeline_test_to_skip(
self, pipeline_test_casse_name, config_class, model_architecture, tokenizer_name, processor_name
):
if pipeline_test_casse_name == "QAPipelineTests":
return True
return False
def setUp(self):
self.model_tester = ErnieMModelTester(self)
self.config_tester = ConfigTester(self, config_class=ErnieMConfig, hidden_size=37)

9
tests/models/oneformer/test_modeling_oneformer.py
View File

@@ -231,6 +231,15 @@ class OneFormerModelTest(ModelTesterMixin, PipelineTesterMixin, unittest.TestCas
test_head_masking = False
test_missing_keys = False
# TODO: Fix the failed tests when this model gets more usage
def is_pipeline_test_to_skip(
self, pipeline_test_casse_name, config_class, model_architecture, tokenizer_name, processor_name
):
if pipeline_test_casse_name == "FeatureExtractionPipelineTests":
return True
return False
def setUp(self):
self.model_tester = OneFormerModelTester(self)
self.config_tester = ConfigTester(self, config_class=OneFormerConfig, has_text_modality=False)

9
tests/models/splinter/test_modeling_splinter.py
View File

@@ -224,6 +224,15 @@ class SplinterModelTest(ModelTesterMixin, PipelineTesterMixin, unittest.TestCase
else {}
)
# TODO: Fix the failed tests when this model gets more usage
def is_pipeline_test_to_skip(
self, pipeline_test_casse_name, config_class, model_architecture, tokenizer_name, processor_name
):
if pipeline_test_casse_name == "QAPipelineTests":
return True
return False
def _prepare_for_class(self, inputs_dict, model_class, return_labels=False):
inputs_dict = copy.deepcopy(inputs_dict)
if return_labels:

18
tests/test_pipeline_mixin.py
View File

@@ -142,18 +142,22 @@ class PipelineTesterMixin:
tokenizer_names = []
processor_names = []
commit = None
if model_arch_name in tiny_model_summary:
tokenizer_names = tiny_model_summary[model_arch_name]["tokenizer_classes"]
processor_names = tiny_model_summary[model_arch_name]["processor_classes"]
commit = tiny_model_summary[model_arch_name]["sha"]
# Adding `None` (if empty) so we can generate tests
tokenizer_names = [None] if len(tokenizer_names) == 0 else tokenizer_names
processor_names = [None] if len(processor_names) == 0 else processor_names
repo_name = f"tiny-random-{model_arch_name}"
self.run_model_pipeline_tests(task, repo_name, model_architecture, tokenizer_names, processor_names)
self.run_model_pipeline_tests(
task, repo_name, model_architecture, tokenizer_names, processor_names, commit
)
def run_model_pipeline_tests(self, task, repo_name, model_architecture, tokenizer_names, processor_names):
def run_model_pipeline_tests(self, task, repo_name, model_architecture, tokenizer_names, processor_names, commit):
"""Run pipeline tests for a specific `task` with the give model class and tokenizer/processor class names
Args:
@@ -187,9 +191,9 @@ class PipelineTesterMixin:
f"`{tokenizer_name}` | processor `{processor_name}`."
)
continue
self.run_pipeline_test(task, repo_name, model_architecture, tokenizer_name, processor_name)
self.run_pipeline_test(task, repo_name, model_architecture, tokenizer_name, processor_name, commit)
def run_pipeline_test(self, task, repo_name, model_architecture, tokenizer_name, processor_name):
def run_pipeline_test(self, task, repo_name, model_architecture, tokenizer_name, processor_name, commit):
"""Run pipeline tests for a specific `task` with the give model class and tokenizer/processor class name
The model will be loaded from a model repository on the Hub.
@@ -211,14 +215,14 @@ class PipelineTesterMixin:
tokenizer = None
if tokenizer_name is not None:
tokenizer_class = getattr(transformers_module, tokenizer_name)
tokenizer = tokenizer_class.from_pretrained(repo_id)
tokenizer = tokenizer_class.from_pretrained(repo_id, revision=commit)
processor = None
if processor_name is not None:
processor_class = getattr(transformers_module, processor_name)
# If the required packages (like `Pillow` or `torchaudio`) are not installed, this will fail.
try:
processor = processor_class.from_pretrained(repo_id)
processor = processor_class.from_pretrained(repo_id, revision=commit)
except Exception:
logger.warning(
f"{self.__class__.__name__}::test_pipeline_{task.replace('-', '_')} is skipped: Could not load the "
@@ -236,7 +240,7 @@ class PipelineTesterMixin:
# TODO: We should check if a model file is on the Hub repo. instead.
try:
model = model_architecture.from_pretrained(repo_id)
model = model_architecture.from_pretrained(repo_id, revision=commit)
except Exception:
logger.warning(
f"{self.__class__.__name__}::test_pipeline_{task.replace('-', '_')} is skipped: Could not find or load "

6068
tests/utils/tiny_model_summary.json
View File

File diff suppressed because it is too large Load Diff