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

[tests] remove TF tests (uses of require_tf) (#38944)

Browse Source 
* remove uses of require_tf

* remove redundant import guards

* this class has no tests

* nits

* del tf rng comment
This commit is contained in:
Joao Gante
2025-06-25 18:29:10 +01:00
committed by GitHub
parent d37f751797
commit 1d45d90e5d
44 changed files with 21 additions and 2504 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
tests/pipelines/test_pipelines_audio_classification.py
View File

@@ -28,7 +28,6 @@ from transformers.testing_utils import (
compare_pipeline_output_to_hub_spec,
is_pipeline_test,
nested_simplify,
require_tf,
require_torch,
require_torchaudio,
slow,
@@ -193,11 +192,6 @@ class AudioClassificationPipelineTests(unittest.TestCase):
],
)
@require_tf
@unittest.skip(reason="Audio classification is not implemented for TF")
def test_small_model_tf(self):
pass
@require_torch
@slow
def test_top_k_none_returns_all_labels(self):

5
tests/pipelines/test_pipelines_automatic_speech_recognition.py
View File

@@ -40,7 +40,6 @@ from transformers.testing_utils import (
is_torch_available,
nested_simplify,
require_pyctcdecode,
require_tf,
require_torch,
require_torch_accelerator,
require_torchaudio,
@@ -326,10 +325,6 @@ class AutomaticSpeechRecognitionPipelineTests(unittest.TestCase):
):
_ = speech_recognizer(filename, return_timestamps="char")
@require_tf
def test_small_model_tf(self):
self.skipTest(reason="Tensorflow not supported yet.")
@require_torch
@unittest.skip("TODO (joao, eustache): this test is failing, find the breaking PR and fix the cause or the test")
def test_torch_small_no_tokenizer_files(self):

76
tests/pipelines/test_pipelines_common.py
View File

@@ -48,8 +48,6 @@ from transformers.testing_utils import (
is_pipeline_test,
is_staging_test,
nested_simplify,
require_tensorflow_probability,
require_tf,
require_torch,
require_torch_accelerator,
require_torch_multi_accelerator,
@@ -177,20 +175,6 @@ class CommonPipelineTest(unittest.TestCase):
results.append(out)
self.assertEqual(len(results), 10)
@require_tf
def test_iterator_data_tf(self):
def data(n: int):
for _ in range(n):
yield "This is a test"
pipe = pipeline(model="hf-internal-testing/tiny-random-distilbert", framework="tf")
out = pipe("This is a test")
results = []
for out in pipe(data(10)):
self.assertEqual(nested_simplify(out), {"label": "LABEL_0", "score": 0.504})
results.append(out)
self.assertEqual(len(results), 10)
@require_torch
def test_unbatch_attentions_hidden_states(self):
model = DistilBertForSequenceClassification.from_pretrained(
@@ -262,9 +246,9 @@ class CommonPipelineTest(unittest.TestCase):
@is_pipeline_test
@require_torch
class PipelineScikitCompatTest(unittest.TestCase):
@require_torch
def test_pipeline_predict_pt(self):
def test_pipeline_predict(self):
data = ["This is a test"]
text_classifier = pipeline(
@@ -275,20 +259,7 @@ class PipelineScikitCompatTest(unittest.TestCase):
actual_output = text_classifier.predict(data)
self.assertEqual(expected_output, actual_output)
@require_tf
def test_pipeline_predict_tf(self):
data = ["This is a test"]
text_classifier = pipeline(
task="text-classification", model="hf-internal-testing/tiny-random-distilbert", framework="tf"
)
expected_output = [{"label": ANY(str), "score": ANY(float)}]
actual_output = text_classifier.predict(data)
self.assertEqual(expected_output, actual_output)
@require_torch
def test_pipeline_transform_pt(self):
def test_pipeline_transform(self):
data = ["This is a test"]
text_classifier = pipeline(
@@ -299,18 +270,6 @@ class PipelineScikitCompatTest(unittest.TestCase):
actual_output = text_classifier.transform(data)
self.assertEqual(expected_output, actual_output)
@require_tf
def test_pipeline_transform_tf(self):
data = ["This is a test"]
text_classifier = pipeline(
task="text-classification", model="hf-internal-testing/tiny-random-distilbert", framework="tf"
)
expected_output = [{"label": ANY(str), "score": ANY(float)}]
actual_output = text_classifier.transform(data)
self.assertEqual(expected_output, actual_output)
@is_pipeline_test
class PipelinePadTest(unittest.TestCase):
@@ -620,23 +579,6 @@ class PipelineUtilsTest(unittest.TestCase):
gc.collect()
backend_empty_cache(torch_device)
@slow
@require_tf
def test_load_default_pipelines_tf(self):
from transformers.modeling_tf_utils import keras
from transformers.pipelines import SUPPORTED_TASKS
set_seed_fn = lambda: keras.utils.set_random_seed(0) # noqa: E731
for task in SUPPORTED_TASKS.keys():
if task == "table-question-answering":
# test table in separate test due to more dependencies
continue
self.check_default_pipeline(task, "tf", set_seed_fn, self.check_models_equal_tf)
# clean-up as much as possible GPU memory occupied by TF
gc.collect()
@slow
@require_torch
def test_load_default_pipelines_pt_table_qa(self):
@@ -663,18 +605,6 @@ class PipelineUtilsTest(unittest.TestCase):
pipe = pipeline("text-generation", device=torch_device)
_ = pipe("Hello")
@slow
@require_tf
@require_tensorflow_probability
def test_load_default_pipelines_tf_table_qa(self):
import tensorflow as tf
set_seed_fn = lambda: tf.random.set_seed(0) # noqa: E731
self.check_default_pipeline("table-question-answering", "tf", set_seed_fn, self.check_models_equal_tf)
# clean-up as much as possible GPU memory occupied by PyTorch
gc.collect()
def check_default_pipeline(self, task, framework, set_seed_fn, check_models_equal_fn):
from transformers.pipelines import SUPPORTED_TASKS, pipeline

6
tests/pipelines/test_pipelines_depth_estimation.py
View File

@@ -24,7 +24,6 @@ from transformers.testing_utils import (
compare_pipeline_output_to_hub_spec,
is_pipeline_test,
nested_simplify,
require_tf,
require_timm,
require_torch,
require_vision,
@@ -123,11 +122,6 @@ class DepthEstimationPipelineTests(unittest.TestCase):
for single_output in outputs:
compare_pipeline_output_to_hub_spec(single_output, DepthEstimationOutput)
@require_tf
@unittest.skip(reason="Depth estimation is not implemented in TF")
def test_small_model_tf(self):
pass
@slow
@require_torch
def test_large_model_pt(self):

6
tests/pipelines/test_pipelines_document_question_answering.py
View File

@@ -27,7 +27,6 @@ from transformers.testing_utils import (
nested_simplify,
require_detectron2,
require_pytesseract,
require_tf,
require_torch,
require_torch_bf16,
require_vision,
@@ -423,8 +422,3 @@ class DocumentQuestionAnsweringPipelineTests(unittest.TestCase):
question = "What is the invoice number?"
outputs = dqa_pipeline(image=image, question=question, top_k=2)
self.assertEqual(nested_simplify(outputs, decimals=4), [{"answer": "us-001"}])
@require_tf
@unittest.skip(reason="Document question answering not implemented in TF")
def test_small_model_tf(self):
pass

64
tests/pipelines/test_pipelines_feature_extraction.py
View File

@@ -23,19 +23,15 @@ from transformers import (
TF_MODEL_MAPPING,
FeatureExtractionPipeline,
LxmertConfig,
is_tf_available,
is_torch_available,
pipeline,
)
from transformers.testing_utils import is_pipeline_test, nested_simplify, require_tf, require_torch
from transformers.testing_utils import is_pipeline_test, nested_simplify, require_torch
if is_torch_available():
import torch
if is_tf_available():
import tensorflow as tf
@is_pipeline_test
class FeatureExtractionPipelineTests(unittest.TestCase):
@@ -52,16 +48,6 @@ class FeatureExtractionPipelineTests(unittest.TestCase):
nested_simplify(outputs),
[[[2.287, 1.234, 0.042, 1.53, 1.306, 0.879, -0.526, -1.71, -1.276, 0.756, -0.775, -1.048, -0.25, -0.595, -0.137, -0.598, 2.022, -0.812, 0.284, -0.488, -0.391, -0.403, -0.525, -0.061, -0.228, 1.086, 0.378, -0.14, 0.599, -0.087, -2.259, -0.098], [1.676, 0.232, -1.508, -0.145, 1.798, -1.388, 1.331, -0.37, -0.939, 0.043, 0.06, -0.414, -1.408, 0.24, 0.622, -0.55, -0.569, 1.873, -0.706, 1.924, -0.254, 1.927, -0.423, 0.152, -0.952, 0.509, -0.496, -0.968, 0.093, -1.049, -0.65, 0.312], [0.207, -0.775, -1.822, 0.321, -0.71, -0.201, 0.3, 1.146, -0.233, -0.753, -0.305, 1.309, -1.47, -0.21, 1.802, -1.555, -1.175, 1.323, -0.303, 0.722, -0.076, 0.103, -1.406, 1.931, 0.091, 0.237, 1.172, 1.607, 0.253, -0.9, -1.068, 0.438], [0.615, 1.077, 0.171, -0.175, 1.3, 0.901, -0.653, -0.138, 0.341, -0.654, -0.184, -0.441, -0.424, 0.356, -0.075, 0.26, -1.023, 0.814, 0.524, -0.904, -0.204, -0.623, 1.234, -1.03, 2.594, 0.56, 1.831, -0.199, -1.508, -0.492, -1.687, -2.165], [0.129, 0.008, -1.279, -0.412, -0.004, 1.663, 0.196, 0.104, 0.123, 0.119, 0.635, 1.757, 2.334, -0.799, -1.626, -1.26, 0.595, -0.316, -1.399, 0.232, 0.264, 1.386, -1.171, -0.256, -0.256, -1.944, 1.168, -0.368, -0.714, -0.51, 0.454, 1.148], [-0.32, 0.29, -1.309, -0.177, 0.453, 0.636, -0.024, 0.509, 0.931, -1.754, -1.575, 0.786, 0.046, -1.165, -1.416, 1.373, 1.293, -0.285, -1.541, -1.186, -0.106, -0.994, 2.001, 0.972, -0.02, 1.654, -0.236, 0.643, 1.02, 0.572, -0.914, -0.154], [0.7, -0.937, 0.441, 0.25, 0.78, -0.022, 0.282, -0.095, 1.558, -0.336, 1.706, 0.884, 1.28, 0.198, -0.796, 1.218, -1.769, 1.197, -0.342, -0.177, -0.645, 1.364, 0.008, -0.597, -0.484, -2.772, -0.696, -0.632, -0.34, -1.527, -0.562, 0.862], [2.504, 0.831, -1.271, -0.033, 0.298, -0.735, 1.339, 1.74, 0.233, -1.424, -0.819, -0.761, 0.291, 0.853, -0.092, -0.885, 0.164, 1.025, 0.907, 0.749, -1.515, -0.545, -1.365, 0.271, 0.034, -2.005, 0.031, 0.244, 0.621, 0.176, 0.336, -1.196], [-0.711, 0.591, -1.001, -0.946, 0.784, -1.66, 1.545, 0.799, -0.857, 1.148, 0.213, -0.285, 0.464, -0.139, 0.79, -1.663, -1.121, 0.575, -0.178, -0.508, 1.565, -0.242, -0.346, 1.024, -1.135, -0.158, -2.101, 0.275, 2.009, -0.425, 0.716, 0.981], [0.912, -1.186, -0.846, -0.421, -1.315, -0.827, 0.309, 0.533, 1.029, -2.343, 1.513, -1.238, 1.487, -0.849, 0.896, -0.927, -0.459, 0.159, 0.177, 0.873, 0.935, 1.433, -0.485, 0.737, 1.327, -0.338, 1.608, -0.47, -0.445, -1.118, -0.213, -0.446], [-0.434, -1.362, -1.098, -1.068, 1.507, 0.003, 0.413, -0.395, 0.897, -0.237, 1.405, -0.344, 1.693, 0.677, 0.097, -0.257, -0.602, 1.026, -1.229, 0.855, -0.713, 1.014, 0.443, 0.238, 0.425, -2.184, 1.933, -1.157, -1.132, -0.597, -0.785, 0.967], [0.58, -0.971, 0.789, -0.468, -0.576, 1.779, 1.747, 1.715, -1.939, 0.125, 0.656, -0.042, -1.024, -1.767, 0.107, -0.408, -0.866, -1.774, 1.248, 0.939, -0.033, 1.523, 1.168, -0.744, 0.209, -0.168, -0.316, 0.207, -0.432, 0.047, -0.646, -0.664], [-0.185, -0.613, -1.695, 1.602, -0.32, -0.277, 0.967, 0.728, -0.965, -0.234, 1.069, -0.63, -1.631, 0.711, 0.426, 1.298, -0.191, -0.467, -0.771, 0.971, -0.118, -1.577, -2.064, -0.055, -0.59, 0.642, -0.997, 1.251, 0.538, 1.367, 0.106, 1.704]]]) # fmt: skip
@require_tf
def test_small_model_tf(self):
feature_extractor = pipeline(
task="feature-extraction", model="hf-internal-testing/tiny-random-distilbert", framework="tf"
)
outputs = feature_extractor("This is a test")
self.assertEqual(
nested_simplify(outputs),
[[[2.287, 1.234, 0.042, 1.53, 1.306, 0.879, -0.526, -1.71, -1.276, 0.756, -0.775, -1.048, -0.25, -0.595, -0.137, -0.598, 2.022, -0.812, 0.284, -0.488, -0.391, -0.403, -0.525, -0.061, -0.228, 1.086, 0.378, -0.14, 0.599, -0.087, -2.259, -0.098], [1.676, 0.232, -1.508, -0.145, 1.798, -1.388, 1.331, -0.37, -0.939, 0.043, 0.06, -0.414, -1.408, 0.24, 0.622, -0.55, -0.569, 1.873, -0.706, 1.924, -0.254, 1.927, -0.423, 0.152, -0.952, 0.509, -0.496, -0.968, 0.093, -1.049, -0.65, 0.312], [0.207, -0.775, -1.822, 0.321, -0.71, -0.201, 0.3, 1.146, -0.233, -0.753, -0.305, 1.309, -1.47, -0.21, 1.802, -1.555, -1.175, 1.323, -0.303, 0.722, -0.076, 0.103, -1.406, 1.931, 0.091, 0.237, 1.172, 1.607, 0.253, -0.9, -1.068, 0.438], [0.615, 1.077, 0.171, -0.175, 1.3, 0.901, -0.653, -0.138, 0.341, -0.654, -0.184, -0.441, -0.424, 0.356, -0.075, 0.26, -1.023, 0.814, 0.524, -0.904, -0.204, -0.623, 1.234, -1.03, 2.594, 0.56, 1.831, -0.199, -1.508, -0.492, -1.687, -2.165], [0.129, 0.008, -1.279, -0.412, -0.004, 1.663, 0.196, 0.104, 0.123, 0.119, 0.635, 1.757, 2.334, -0.799, -1.626, -1.26, 0.595, -0.316, -1.399, 0.232, 0.264, 1.386, -1.171, -0.256, -0.256, -1.944, 1.168, -0.368, -0.714, -0.51, 0.454, 1.148], [-0.32, 0.29, -1.309, -0.177, 0.453, 0.636, -0.024, 0.509, 0.931, -1.754, -1.575, 0.786, 0.046, -1.165, -1.416, 1.373, 1.293, -0.285, -1.541, -1.186, -0.106, -0.994, 2.001, 0.972, -0.02, 1.654, -0.236, 0.643, 1.02, 0.572, -0.914, -0.154], [0.7, -0.937, 0.441, 0.25, 0.78, -0.022, 0.282, -0.095, 1.558, -0.336, 1.706, 0.884, 1.28, 0.198, -0.796, 1.218, -1.769, 1.197, -0.342, -0.177, -0.645, 1.364, 0.008, -0.597, -0.484, -2.772, -0.696, -0.632, -0.34, -1.527, -0.562, 0.862], [2.504, 0.831, -1.271, -0.033, 0.298, -0.735, 1.339, 1.74, 0.233, -1.424, -0.819, -0.761, 0.291, 0.853, -0.092, -0.885, 0.164, 1.025, 0.907, 0.749, -1.515, -0.545, -1.365, 0.271, 0.034, -2.005, 0.031, 0.244, 0.621, 0.176, 0.336, -1.196], [-0.711, 0.591, -1.001, -0.946, 0.784, -1.66, 1.545, 0.799, -0.857, 1.148, 0.213, -0.285, 0.464, -0.139, 0.79, -1.663, -1.121, 0.575, -0.178, -0.508, 1.565, -0.242, -0.346, 1.024, -1.135, -0.158, -2.101, 0.275, 2.009, -0.425, 0.716, 0.981], [0.912, -1.186, -0.846, -0.421, -1.315, -0.827, 0.309, 0.533, 1.029, -2.343, 1.513, -1.238, 1.487, -0.849, 0.896, -0.927, -0.459, 0.159, 0.177, 0.873, 0.935, 1.433, -0.485, 0.737, 1.327, -0.338, 1.608, -0.47, -0.445, -1.118, -0.213, -0.446], [-0.434, -1.362, -1.098, -1.068, 1.507, 0.003, 0.413, -0.395, 0.897, -0.237, 1.405, -0.344, 1.693, 0.677, 0.097, -0.257, -0.602, 1.026, -1.229, 0.855, -0.713, 1.014, 0.443, 0.238, 0.425, -2.184, 1.933, -1.157, -1.132, -0.597, -0.785, 0.967], [0.58, -0.971, 0.789, -0.468, -0.576, 1.779, 1.747, 1.715, -1.939, 0.125, 0.656, -0.042, -1.024, -1.767, 0.107, -0.408, -0.866, -1.774, 1.248, 0.939, -0.033, 1.523, 1.168, -0.744, 0.209, -0.168, -0.316, 0.207, -0.432, 0.047, -0.646, -0.664], [-0.185, -0.613, -1.695, 1.602, -0.32, -0.277, 0.967, 0.728, -0.965, -0.234, 1.069, -0.63, -1.631, 0.711, 0.426, 1.298, -0.191, -0.467, -0.771, 0.971, -0.118, -1.577, -2.064, -0.055, -0.59, 0.642, -0.997, 1.251, 0.538, 1.367, 0.106, 1.704]]]) # fmt: skip
@require_torch
def test_tokenization_small_model_pt(self):
feature_extractor = pipeline(
@@ -102,46 +88,6 @@ class FeatureExtractionPipelineTests(unittest.TestCase):
tokenize_kwargs=tokenize_kwargs,
)
@require_tf
def test_tokenization_small_model_tf(self):
feature_extractor = pipeline(
task="feature-extraction", model="hf-internal-testing/tiny-random-distilbert", framework="tf"
)
# test with empty parameters
outputs = feature_extractor("This is a test")
self.assertEqual(
nested_simplify(outputs),
[[[2.287, 1.234, 0.042, 1.53, 1.306, 0.879, -0.526, -1.71, -1.276, 0.756, -0.775, -1.048, -0.25, -0.595, -0.137, -0.598, 2.022, -0.812, 0.284, -0.488, -0.391, -0.403, -0.525, -0.061, -0.228, 1.086, 0.378, -0.14, 0.599, -0.087, -2.259, -0.098], [1.676, 0.232, -1.508, -0.145, 1.798, -1.388, 1.331, -0.37, -0.939, 0.043, 0.06, -0.414, -1.408, 0.24, 0.622, -0.55, -0.569, 1.873, -0.706, 1.924, -0.254, 1.927, -0.423, 0.152, -0.952, 0.509, -0.496, -0.968, 0.093, -1.049, -0.65, 0.312], [0.207, -0.775, -1.822, 0.321, -0.71, -0.201, 0.3, 1.146, -0.233, -0.753, -0.305, 1.309, -1.47, -0.21, 1.802, -1.555, -1.175, 1.323, -0.303, 0.722, -0.076, 0.103, -1.406, 1.931, 0.091, 0.237, 1.172, 1.607, 0.253, -0.9, -1.068, 0.438], [0.615, 1.077, 0.171, -0.175, 1.3, 0.901, -0.653, -0.138, 0.341, -0.654, -0.184, -0.441, -0.424, 0.356, -0.075, 0.26, -1.023, 0.814, 0.524, -0.904, -0.204, -0.623, 1.234, -1.03, 2.594, 0.56, 1.831, -0.199, -1.508, -0.492, -1.687, -2.165], [0.129, 0.008, -1.279, -0.412, -0.004, 1.663, 0.196, 0.104, 0.123, 0.119, 0.635, 1.757, 2.334, -0.799, -1.626, -1.26, 0.595, -0.316, -1.399, 0.232, 0.264, 1.386, -1.171, -0.256, -0.256, -1.944, 1.168, -0.368, -0.714, -0.51, 0.454, 1.148], [-0.32, 0.29, -1.309, -0.177, 0.453, 0.636, -0.024, 0.509, 0.931, -1.754, -1.575, 0.786, 0.046, -1.165, -1.416, 1.373, 1.293, -0.285, -1.541, -1.186, -0.106, -0.994, 2.001, 0.972, -0.02, 1.654, -0.236, 0.643, 1.02, 0.572, -0.914, -0.154], [0.7, -0.937, 0.441, 0.25, 0.78, -0.022, 0.282, -0.095, 1.558, -0.336, 1.706, 0.884, 1.28, 0.198, -0.796, 1.218, -1.769, 1.197, -0.342, -0.177, -0.645, 1.364, 0.008, -0.597, -0.484, -2.772, -0.696, -0.632, -0.34, -1.527, -0.562, 0.862], [2.504, 0.831, -1.271, -0.033, 0.298, -0.735, 1.339, 1.74, 0.233, -1.424, -0.819, -0.761, 0.291, 0.853, -0.092, -0.885, 0.164, 1.025, 0.907, 0.749, -1.515, -0.545, -1.365, 0.271, 0.034, -2.005, 0.031, 0.244, 0.621, 0.176, 0.336, -1.196], [-0.711, 0.591, -1.001, -0.946, 0.784, -1.66, 1.545, 0.799, -0.857, 1.148, 0.213, -0.285, 0.464, -0.139, 0.79, -1.663, -1.121, 0.575, -0.178, -0.508, 1.565, -0.242, -0.346, 1.024, -1.135, -0.158, -2.101, 0.275, 2.009, -0.425, 0.716, 0.981], [0.912, -1.186, -0.846, -0.421, -1.315, -0.827, 0.309, 0.533, 1.029, -2.343, 1.513, -1.238, 1.487, -0.849, 0.896, -0.927, -0.459, 0.159, 0.177, 0.873, 0.935, 1.433, -0.485, 0.737, 1.327, -0.338, 1.608, -0.47, -0.445, -1.118, -0.213, -0.446], [-0.434, -1.362, -1.098, -1.068, 1.507, 0.003, 0.413, -0.395, 0.897, -0.237, 1.405, -0.344, 1.693, 0.677, 0.097, -0.257, -0.602, 1.026, -1.229, 0.855, -0.713, 1.014, 0.443, 0.238, 0.425, -2.184, 1.933, -1.157, -1.132, -0.597, -0.785, 0.967], [0.58, -0.971, 0.789, -0.468, -0.576, 1.779, 1.747, 1.715, -1.939, 0.125, 0.656, -0.042, -1.024, -1.767, 0.107, -0.408, -0.866, -1.774, 1.248, 0.939, -0.033, 1.523, 1.168, -0.744, 0.209, -0.168, -0.316, 0.207, -0.432, 0.047, -0.646, -0.664], [-0.185, -0.613, -1.695, 1.602, -0.32, -0.277, 0.967, 0.728, -0.965, -0.234, 1.069, -0.63, -1.631, 0.711, 0.426, 1.298, -0.191, -0.467, -0.771, 0.971, -0.118, -1.577, -2.064, -0.055, -0.59, 0.642, -0.997, 1.251, 0.538, 1.367, 0.106, 1.704]]]) # fmt: skip
# test with various tokenizer parameters
tokenize_kwargs = {"max_length": 3}
outputs = feature_extractor("This is a test", tokenize_kwargs=tokenize_kwargs)
self.assertEqual(np.squeeze(outputs).shape, (3, 32))
tokenize_kwargs = {"truncation": True, "padding": True, "max_length": 4}
outputs = feature_extractor(
["This is a test", "This", "This is", "This is a", "This is a test test test test"],
tokenize_kwargs=tokenize_kwargs,
)
self.assertEqual(np.squeeze(outputs).shape, (5, 4, 32))
tokenize_kwargs = {"padding": True, "max_length": 4}
outputs = feature_extractor(
["This is a test", "This", "This is", "This is a", "This is a test test test test"],
truncation=True,
tokenize_kwargs=tokenize_kwargs,
)
self.assertEqual(np.squeeze(outputs).shape, (5, 4, 32))
# raise value error if truncation parameter given for two places
tokenize_kwargs = {"truncation": True}
with self.assertRaises(ValueError):
_ = feature_extractor(
["This is a test", "This", "This is", "This is a", "This is a test test test test"],
truncation=True,
tokenize_kwargs=tokenize_kwargs,
)
@require_torch
def test_return_tensors_pt(self):
feature_extractor = pipeline(
@@ -150,14 +96,6 @@ class FeatureExtractionPipelineTests(unittest.TestCase):
outputs = feature_extractor("This is a test", return_tensors=True)
self.assertTrue(torch.is_tensor(outputs))
@require_tf
def test_return_tensors_tf(self):
feature_extractor = pipeline(
task="feature-extraction", model="hf-internal-testing/tiny-random-distilbert", framework="tf"
)
outputs = feature_extractor("This is a test", return_tensors=True)
self.assertTrue(tf.is_tensor(outputs))
def get_shape(self, input_, shape=None):
if shape is None:
shape = []

55
tests/pipelines/test_pipelines_fill_mask.py
View File

@@ -22,7 +22,6 @@ from transformers.testing_utils import (
is_pipeline_test,
is_torch_available,
nested_simplify,
require_tf,
require_torch,
require_torch_accelerator,
slow,
@@ -44,47 +43,6 @@ class FillMaskPipelineTests(unittest.TestCase):
if is_torch_available():
backend_empty_cache(torch_device)
@require_tf
def test_small_model_tf(self):
unmasker = pipeline(task="fill-mask", model="sshleifer/tiny-distilroberta-base", top_k=2, framework="tf")
outputs = unmasker("My name is <mask>")
self.assertEqual(
nested_simplify(outputs, decimals=6),
[
{"sequence": "My name is grouped", "score": 2.1e-05, "token": 38015, "token_str": " grouped"},
{"sequence": "My name is accuser", "score": 2.1e-05, "token": 25506, "token_str": " accuser"},
],
)
outputs = unmasker("The largest city in France is <mask>")
self.assertEqual(
nested_simplify(outputs, decimals=6),
[
{
"sequence": "The largest city in France is grouped",
"score": 2.1e-05,
"token": 38015,
"token_str": " grouped",
},
{
"sequence": "The largest city in France is accuser",
"score": 2.1e-05,
"token": 25506,
"token_str": " accuser",
},
],
)
outputs = unmasker("My name is <mask>", targets=[" Patrick", " Clara", " Teven"], top_k=3)
self.assertEqual(
nested_simplify(outputs, decimals=6),
[
{"sequence": "My name is Clara", "score": 2e-05, "token": 13606, "token_str": " Clara"},
{"sequence": "My name is Patrick", "score": 2e-05, "token": 3499, "token_str": " Patrick"},
{"sequence": "My name is Te", "score": 1.9e-05, "token": 2941, "token_str": " Te"},
],
)
@require_torch
def test_small_model_pt(self):
unmasker = pipeline(task="fill-mask", model="sshleifer/tiny-distilroberta-base", top_k=2, framework="pt")
@@ -172,12 +130,6 @@ class FillMaskPipelineTests(unittest.TestCase):
unmasker = pipeline(task="fill-mask", model="distilbert/distilroberta-base", top_k=2, framework="pt")
self.run_large_test(unmasker)
@slow
@require_tf
def test_large_model_tf(self):
unmasker = pipeline(task="fill-mask", model="distilbert/distilroberta-base", top_k=2, framework="tf")
self.run_large_test(unmasker)
def run_large_test(self, unmasker):
outputs = unmasker("My name is <mask>")
self.assertEqual(
@@ -244,13 +196,6 @@ class FillMaskPipelineTests(unittest.TestCase):
unmasker.tokenizer.pad_token = None
self.run_pipeline_test(unmasker, [])
@require_tf
def test_model_no_pad_tf(self):
unmasker = pipeline(task="fill-mask", model="sshleifer/tiny-distilroberta-base", framework="tf")
unmasker.tokenizer.pad_token_id = None
unmasker.tokenizer.pad_token = None
self.run_pipeline_test(unmasker, [])
def get_test_pipeline(
self,
model,

27
tests/pipelines/test_pipelines_image_classification.py
View File

@@ -29,7 +29,6 @@ from transformers.testing_utils import (
compare_pipeline_output_to_hub_spec,
is_pipeline_test,
nested_simplify,
require_tf,
require_torch,
require_torch_or_tf,
require_vision,
@@ -175,32 +174,6 @@ class ImageClassificationPipelineTests(unittest.TestCase):
],
)
@require_tf
def test_small_model_tf(self):
small_model = "hf-internal-testing/tiny-random-vit"
image_classifier = pipeline("image-classification", model=small_model, framework="tf")
outputs = image_classifier("http://images.cocodataset.org/val2017/000000039769.jpg")
self.assertEqual(
nested_simplify(outputs, decimals=4),
[{"label": "LABEL_1", "score": 0.574}, {"label": "LABEL_0", "score": 0.426}],
)
outputs = image_classifier(
[
"http://images.cocodataset.org/val2017/000000039769.jpg",
"http://images.cocodataset.org/val2017/000000039769.jpg",
],
top_k=2,
)
self.assertEqual(
nested_simplify(outputs, decimals=4),
[
[{"label": "LABEL_1", "score": 0.574}, {"label": "LABEL_0", "score": 0.426}],
[{"label": "LABEL_1", "score": 0.574}, {"label": "LABEL_0", "score": 0.426}],
],
)
def test_custom_tokenizer(self):
tokenizer = PreTrainedTokenizerBase()

59
tests/pipelines/test_pipelines_image_feature_extraction.py
View File

@@ -22,20 +22,16 @@ from transformers import (
TF_MODEL_MAPPING,
TOKENIZER_MAPPING,
ImageFeatureExtractionPipeline,
is_tf_available,
is_torch_available,
is_vision_available,
pipeline,
)
from transformers.testing_utils import is_pipeline_test, nested_simplify, require_tf, require_torch
from transformers.testing_utils import is_pipeline_test, nested_simplify, require_torch
if is_torch_available():
import torch
if is_tf_available():
import tensorflow as tf
if is_vision_available():
from PIL import Image
@@ -73,28 +69,6 @@ class ImageFeatureExtractionPipelineTests(unittest.TestCase):
nested_simplify(outputs[0]),
[-0.056, 0.083, 0.021, 0.038, 0.242, -0.279, -0.033, -0.003, 0.200, -0.192, 0.045, -0.095, -0.077, 0.017, -0.058, -0.063, -0.029, -0.204, 0.014, 0.042, 0.305, -0.205, -0.099, 0.146, -0.287, 0.020, 0.168, -0.052, 0.046, 0.048, -0.156, 0.093]) # fmt: skip
@require_tf
def test_small_model_tf(self):
feature_extractor = pipeline(
task="image-feature-extraction", model="hf-internal-testing/tiny-random-vit-w-pooler", framework="tf"
)
img = prepare_img()
outputs = feature_extractor(img)
self.assertEqual(
nested_simplify(outputs[0][0]),
[-1.417, -0.392, -1.264, -1.196, 1.648, 0.885, 0.56, -0.606, -1.175, 0.823, 1.912, 0.081, -0.053, 1.119, -0.062, -1.757, -0.571, 0.075, 0.959, 0.118, 1.201, -0.672, -0.498, 0.364, 0.937, -1.623, 0.228, 0.19, 1.697, -1.115, 0.583, -0.981]) # fmt: skip
@require_tf
def test_small_model_w_pooler_tf(self):
feature_extractor = pipeline(
task="image-feature-extraction", model="hf-internal-testing/tiny-random-vit-w-pooler", framework="tf"
)
img = prepare_img()
outputs = feature_extractor(img, pool=True)
self.assertEqual(
nested_simplify(outputs[0]),
[-0.056, 0.083, 0.021, 0.038, 0.242, -0.279, -0.033, -0.003, 0.200, -0.192, 0.045, -0.095, -0.077, 0.017, -0.058, -0.063, -0.029, -0.204, 0.014, 0.042, 0.305, -0.205, -0.099, 0.146, -0.287, 0.020, 0.168, -0.052, 0.046, 0.048, -0.156, 0.093]) # fmt: skip
@require_torch
def test_image_processing_small_model_pt(self):
feature_extractor = pipeline(
@@ -117,28 +91,6 @@ class ImageFeatureExtractionPipelineTests(unittest.TestCase):
outputs = feature_extractor(img, pool=True)
self.assertEqual(np.squeeze(outputs).shape, (32,))
@require_tf
def test_image_processing_small_model_tf(self):
feature_extractor = pipeline(
task="image-feature-extraction", model="hf-internal-testing/tiny-random-vit", framework="tf"
)
# test with image processor parameters
image_processor_kwargs = {"size": {"height": 300, "width": 300}}
img = prepare_img()
with pytest.raises(ValueError):
# Image doesn't match model input size
feature_extractor(img, image_processor_kwargs=image_processor_kwargs)
image_processor_kwargs = {"image_mean": [0, 0, 0], "image_std": [1, 1, 1]}
img = prepare_img()
outputs = feature_extractor(img, image_processor_kwargs=image_processor_kwargs)
self.assertEqual(np.squeeze(outputs).shape, (226, 32))
# Test pooling option
outputs = feature_extractor(img, pool=True)
self.assertEqual(np.squeeze(outputs).shape, (32,))
@require_torch
def test_return_tensors_pt(self):
feature_extractor = pipeline(
@@ -148,15 +100,6 @@ class ImageFeatureExtractionPipelineTests(unittest.TestCase):
outputs = feature_extractor(img, return_tensors=True)
self.assertTrue(torch.is_tensor(outputs))
@require_tf
def test_return_tensors_tf(self):
feature_extractor = pipeline(
task="image-feature-extraction", model="hf-internal-testing/tiny-random-vit", framework="tf"
)
img = prepare_img()
outputs = feature_extractor(img, return_tensors=True)
self.assertTrue(tf.is_tensor(outputs))
def get_test_pipeline(
self,
model,

6
tests/pipelines/test_pipelines_image_segmentation.py
View File

@@ -39,7 +39,6 @@ from transformers.testing_utils import (
compare_pipeline_output_to_hub_spec,
is_pipeline_test,
nested_simplify,
require_tf,
require_timm,
require_torch,
require_vision,
@@ -202,11 +201,6 @@ class ImageSegmentationPipelineTests(unittest.TestCase):
for output_element in single_output:
compare_pipeline_output_to_hub_spec(output_element, ImageSegmentationOutputElement)
@require_tf
@unittest.skip(reason="Image segmentation not implemented in TF")
def test_small_model_tf(self):
pass
@require_torch
def test_small_model_pt_no_panoptic(self):
model_id = "hf-internal-testing/tiny-random-mobilevit"

6
tests/pipelines/test_pipelines_mask_generation.py
View File

@@ -29,7 +29,6 @@ from transformers.testing_utils import (
Expectations,
is_pipeline_test,
nested_simplify,
require_tf,
require_torch,
require_vision,
slow,
@@ -103,11 +102,6 @@ class MaskGenerationPipelineTests(unittest.TestCase):
def run_pipeline_test(self, mask_generator, examples):
pass
@require_tf
@unittest.skip(reason="Image segmentation not implemented in TF")
def test_small_model_tf(self):
pass
@slow
@require_torch
def test_small_model_pt(self):

6
tests/pipelines/test_pipelines_object_detection.py
View File

@@ -30,7 +30,6 @@ from transformers.testing_utils import (
is_pipeline_test,
nested_simplify,
require_pytesseract,
require_tf,
require_timm,
require_torch,
require_vision,
@@ -128,11 +127,6 @@ class ObjectDetectionPipelineTests(unittest.TestCase):
)
compare_pipeline_output_to_hub_spec(detected_object, ObjectDetectionOutputElement)
@require_tf
@unittest.skip(reason="Object detection not implemented in TF")
def test_small_model_tf(self):
pass
@require_torch
def test_small_model_pt(self):
model_id = "hf-internal-testing/tiny-detr-mobilenetsv3"

22
tests/pipelines/test_pipelines_question_answering.py
View File

@@ -29,7 +29,6 @@ from transformers.testing_utils import (
is_pipeline_test,
is_torch_available,
nested_simplify,
require_tf,
require_torch,
require_torch_or_tf,
slow,
@@ -296,17 +295,6 @@ class QAPipelineTests(unittest.TestCase):
answers = [output["answer"] for output in outputs]
self.assertEqual(len(answers), len(set(answers)), "There are duplicate answers in the outputs.")
@require_tf
def test_small_model_tf(self):
question_answerer = pipeline(
"question-answering", model="sshleifer/tiny-distilbert-base-cased-distilled-squad", framework="tf"
)
outputs = question_answerer(
question="Where was HuggingFace founded ?", context="HuggingFace was founded in Paris."
)
self.assertEqual(nested_simplify(outputs), {"score": 0.011, "start": 0, "end": 11, "answer": "HuggingFace"})
@slow
@require_torch
def test_large_model_pt(self):
@@ -421,16 +409,6 @@ between them. It's straightforward to train your models with one before loading
{"answer": "Jax, PyTorch and TensorFlow", "end": 1919, "score": 0.971, "start": 1892},
)
@slow
@require_tf
def test_large_model_tf(self):
question_answerer = pipeline("question-answering", framework="tf")
outputs = question_answerer(
question="Where was HuggingFace founded ?", context="HuggingFace was founded in Paris."
)
self.assertEqual(nested_simplify(outputs), {"score": 0.979, "start": 27, "end": 32, "answer": "Paris"})
@require_torch_or_tf
class QuestionAnsweringArgumentHandlerTests(unittest.TestCase):

228
tests/pipelines/test_pipelines_table_question_answering.py
View File

@@ -26,7 +26,6 @@ from transformers.testing_utils import (
is_pipeline_test,
require_pandas,
require_tensorflow_probability,
require_tf,
require_torch,
slow,
)
@@ -38,111 +37,6 @@ class TQAPipelineTests(unittest.TestCase):
# which are needed to generate automatic tests
model_mapping = MODEL_FOR_TABLE_QUESTION_ANSWERING_MAPPING
@require_tensorflow_probability
@require_pandas
@require_tf
@require_torch
def test_small_model_tf(self):
model_id = "lysandre/tiny-tapas-random-wtq"
model = TFAutoModelForTableQuestionAnswering.from_pretrained(model_id, from_pt=True)
tokenizer = AutoTokenizer.from_pretrained(model_id)
self.assertIsInstance(model.config.aggregation_labels, dict)
self.assertIsInstance(model.config.no_aggregation_label_index, int)
table_querier = TableQuestionAnsweringPipeline(model=model, tokenizer=tokenizer, max_new_tokens=20)
outputs = table_querier(
table={
"actors": ["brad pitt", "leonardo di caprio", "george clooney"],
"age": ["56", "45", "59"],
"number of movies": ["87", "53", "69"],
"date of birth": ["7 february 1967", "10 june 1996", "28 november 1967"],
},
query="how many movies has george clooney played in?",
)
self.assertEqual(
outputs,
{"answer": "AVERAGE > ", "coordinates": [], "cells": [], "aggregator": "AVERAGE"},
)
outputs = table_querier(
table={
"actors": ["brad pitt", "leonardo di caprio", "george clooney"],
"age": ["56", "45", "59"],
"number of movies": ["87", "53", "69"],
"date of birth": ["7 february 1967", "10 june 1996", "28 november 1967"],
},
query=["how many movies has george clooney played in?", "how old is he?", "what's his date of birth?"],
)
self.assertEqual(
outputs,
[
{"answer": "AVERAGE > ", "coordinates": [], "cells": [], "aggregator": "AVERAGE"},
{"answer": "AVERAGE > ", "coordinates": [], "cells": [], "aggregator": "AVERAGE"},
{"answer": "AVERAGE > ", "coordinates": [], "cells": [], "aggregator": "AVERAGE"},
],
)
outputs = table_querier(
table={
"Repository": ["Transformers", "Datasets", "Tokenizers"],
"Stars": ["36542", "4512", "3934"],
"Contributors": ["651", "77", "34"],
"Programming language": ["Python", "Python", "Rust, Python and NodeJS"],
},
query=[
"What repository has the largest number of stars?",
"Given that the numbers of stars defines if a repository is active, what repository is the most"
" active?",
"What is the number of repositories?",
"What is the average number of stars?",
"What is the total amount of stars?",
],
)
self.assertEqual(
outputs,
[
{"answer": "AVERAGE > ", "coordinates": [], "cells": [], "aggregator": "AVERAGE"},
{"answer": "AVERAGE > ", "coordinates": [], "cells": [], "aggregator": "AVERAGE"},
{"answer": "AVERAGE > ", "coordinates": [], "cells": [], "aggregator": "AVERAGE"},
{"answer": "AVERAGE > ", "coordinates": [], "cells": [], "aggregator": "AVERAGE"},
{"answer": "AVERAGE > ", "coordinates": [], "cells": [], "aggregator": "AVERAGE"},
],
)
with self.assertRaises(ValueError):
table_querier(query="What does it do with empty context ?", table=None)
with self.assertRaises(ValueError):
table_querier(query="What does it do with empty context ?", table="")
with self.assertRaises(ValueError):
table_querier(query="What does it do with empty context ?", table={})
with self.assertRaises(ValueError):
table_querier(
table={
"Repository": ["Transformers", "Datasets", "Tokenizers"],
"Stars": ["36542", "4512", "3934"],
"Contributors": ["651", "77", "34"],
"Programming language": ["Python", "Python", "Rust, Python and NodeJS"],
}
)
with self.assertRaises(ValueError):
table_querier(
query="",
table={
"Repository": ["Transformers", "Datasets", "Tokenizers"],
"Stars": ["36542", "4512", "3934"],
"Contributors": ["651", "77", "34"],
"Programming language": ["Python", "Python", "Rust, Python and NodeJS"],
},
)
with self.assertRaises(ValueError):
table_querier(
query=None,
table={
"Repository": ["Transformers", "Datasets", "Tokenizers"],
"Stars": ["36542", "4512", "3934"],
"Contributors": ["651", "77", "34"],
"Programming language": ["Python", "Python", "Rust, Python and NodeJS"],
},
)
@require_torch
def test_small_model_pt(self, torch_dtype="float32"):
model_id = "lysandre/tiny-tapas-random-wtq"
@@ -372,128 +266,6 @@ class TQAPipelineTests(unittest.TestCase):
def test_slow_tokenizer_sqa_pt_fp16(self):
self.test_slow_tokenizer_sqa_pt(torch_dtype="float16")
@require_tf
@require_tensorflow_probability
@require_pandas
@require_torch
def test_slow_tokenizer_sqa_tf(self):
model_id = "lysandre/tiny-tapas-random-sqa"
model = TFAutoModelForTableQuestionAnswering.from_pretrained(model_id, from_pt=True)
tokenizer = AutoTokenizer.from_pretrained(model_id)
table_querier = TableQuestionAnsweringPipeline(model=model, tokenizer=tokenizer, max_new_tokens=20)
inputs = {
"table": {
"actors": ["brad pitt", "leonardo di caprio", "george clooney"],
"age": ["56", "45", "59"],
"number of movies": ["87", "53", "69"],
"date of birth": ["7 february 1967", "10 june 1996", "28 november 1967"],
},
"query": ["how many movies has george clooney played in?", "how old is he?", "what's his date of birth?"],
}
sequential_outputs = table_querier(**inputs, sequential=True)
batch_outputs = table_querier(**inputs, sequential=False)
self.assertEqual(len(sequential_outputs), 3)
self.assertEqual(len(batch_outputs), 3)
self.assertEqual(sequential_outputs[0], batch_outputs[0])
self.assertNotEqual(sequential_outputs[1], batch_outputs[1])
# self.assertNotEqual(sequential_outputs[2], batch_outputs[2])
table_querier = TableQuestionAnsweringPipeline(model=model, tokenizer=tokenizer, max_new_tokens=20)
outputs = table_querier(
table={
"actors": ["brad pitt", "leonardo di caprio", "george clooney"],
"age": ["56", "45", "59"],
"number of movies": ["87", "53", "69"],
"date of birth": ["7 february 1967", "10 june 1996", "28 november 1967"],
},
query="how many movies has george clooney played in?",
)
self.assertEqual(
outputs,
{"answer": "7 february 1967", "coordinates": [(0, 3)], "cells": ["7 february 1967"]},
)
outputs = table_querier(
table={
"actors": ["brad pitt", "leonardo di caprio", "george clooney"],
"age": ["56", "45", "59"],
"number of movies": ["87", "53", "69"],
"date of birth": ["7 february 1967", "10 june 1996", "28 november 1967"],
},
query=["how many movies has george clooney played in?", "how old is he?", "what's his date of birth?"],
)
self.assertEqual(
outputs,
[
{"answer": "7 february 1967", "coordinates": [(0, 3)], "cells": ["7 february 1967"]},
{"answer": "7 february 1967", "coordinates": [(0, 3)], "cells": ["7 february 1967"]},
{"answer": "7 february 1967", "coordinates": [(0, 3)], "cells": ["7 february 1967"]},
],
)
outputs = table_querier(
table={
"Repository": ["Transformers", "Datasets", "Tokenizers"],
"Stars": ["36542", "4512", "3934"],
"Contributors": ["651", "77", "34"],
"Programming language": ["Python", "Python", "Rust, Python and NodeJS"],
},
query=[
"What repository has the largest number of stars?",
"Given that the numbers of stars defines if a repository is active, what repository is the most"
" active?",
"What is the number of repositories?",
"What is the average number of stars?",
"What is the total amount of stars?",
],
)
self.assertEqual(
outputs,
[
{"answer": "Python, Python", "coordinates": [(0, 3), (1, 3)], "cells": ["Python", "Python"]},
{"answer": "Python, Python", "coordinates": [(0, 3), (1, 3)], "cells": ["Python", "Python"]},
{"answer": "Python, Python", "coordinates": [(0, 3), (1, 3)], "cells": ["Python", "Python"]},
{"answer": "Python, Python", "coordinates": [(0, 3), (1, 3)], "cells": ["Python", "Python"]},
{"answer": "Python, Python", "coordinates": [(0, 3), (1, 3)], "cells": ["Python", "Python"]},
],
)
with self.assertRaises(ValueError):
table_querier(query="What does it do with empty context ?", table=None)
with self.assertRaises(ValueError):
table_querier(query="What does it do with empty context ?", table="")
with self.assertRaises(ValueError):
table_querier(query="What does it do with empty context ?", table={})
with self.assertRaises(ValueError):
table_querier(
table={
"Repository": ["Transformers", "Datasets", "Tokenizers"],
"Stars": ["36542", "4512", "3934"],
"Contributors": ["651", "77", "34"],
"Programming language": ["Python", "Python", "Rust, Python and NodeJS"],
}
)
with self.assertRaises(ValueError):
table_querier(
query="",
table={
"Repository": ["Transformers", "Datasets", "Tokenizers"],
"Stars": ["36542", "4512", "3934"],
"Contributors": ["651", "77", "34"],
"Programming language": ["Python", "Python", "Rust, Python and NodeJS"],
},
)
with self.assertRaises(ValueError):
table_querier(
query=None,
table={
"Repository": ["Transformers", "Datasets", "Tokenizers"],
"Stars": ["36542", "4512", "3934"],
"Contributors": ["651", "77", "34"],
"Programming language": ["Python", "Python", "Rust, Python and NodeJS"],
},
)
@slow
@require_torch
def test_integration_wtq_pt(self, torch_dtype="float32"):

22
tests/pipelines/test_pipelines_text_classification.py
View File

@@ -24,7 +24,6 @@ from transformers.testing_utils import (
is_pipeline_test,
is_torch_available,
nested_simplify,
require_tf,
require_torch,
require_torch_bf16,
require_torch_fp16,
@@ -152,15 +151,6 @@ class TextClassificationPipelineTests(unittest.TestCase):
outputs = text_classifier("This is great !")
self.assertEqual(nested_simplify(outputs), [{"label": "LABEL_0", "score": 0.504}])
@require_tf
def test_small_model_tf(self):
text_classifier = pipeline(
task="text-classification", model="hf-internal-testing/tiny-random-distilbert", framework="tf"
)
outputs = text_classifier("This is great !")
self.assertEqual(nested_simplify(outputs), [{"label": "LABEL_0", "score": 0.504}])
@slow
@require_torch
def test_pt_bert(self):
@@ -173,18 +163,6 @@ class TextClassificationPipelineTests(unittest.TestCase):
outputs = text_classifier("Birds are a type of animal")
self.assertEqual(nested_simplify(outputs), [{"label": "POSITIVE", "score": 0.988}])
@slow
@require_tf
def test_tf_bert(self):
text_classifier = pipeline("text-classification", framework="tf")
outputs = text_classifier("This is great !")
self.assertEqual(nested_simplify(outputs), [{"label": "POSITIVE", "score": 1.0}])
outputs = text_classifier("This is bad !")
self.assertEqual(nested_simplify(outputs), [{"label": "NEGATIVE", "score": 1.0}])
outputs = text_classifier("Birds are a type of animal")
self.assertEqual(nested_simplify(outputs), [{"label": "POSITIVE", "score": 0.988}])
def get_test_pipeline(
self,
model,

21
tests/pipelines/test_pipelines_token_classification.py
View File

@@ -29,7 +29,6 @@ from transformers.testing_utils import (
is_pipeline_test,
is_torch_available,
nested_simplify,
require_tf,
require_torch,
require_torch_accelerator,
slow,
@@ -823,26 +822,6 @@ class TokenClassificationPipelineTests(unittest.TestCase):
[("▁I", False), ("▁play", False), ("▁the", False), ("▁there", False), ("min", True)],
)
@require_tf
def test_tf_only(self):
model_name = "hf-internal-testing/tiny-random-bert-tf-only" # This model only has a TensorFlow version
# We test that if we don't specify framework='tf', it gets detected automatically
token_classifier = pipeline(task="ner", model=model_name)
self.assertEqual(token_classifier.framework, "tf")
@require_tf
def test_small_model_tf(self):
model_name = "hf-internal-testing/tiny-bert-for-token-classification"
token_classifier = pipeline(task="token-classification", model=model_name, framework="tf")
outputs = token_classifier("This is a test !")
self.assertEqual(
nested_simplify(outputs),
[
{"entity": "I-MISC", "score": 0.115, "index": 1, "word": "this", "start": 0, "end": 4},
{"entity": "I-MISC", "score": 0.115, "index": 2, "word": "is", "start": 5, "end": 7},
],
)
@require_torch
def test_no_offset_tokenizer(self):
model_name = "hf-internal-testing/tiny-bert-for-token-classification"

6
tests/pipelines/test_pipelines_video_classification.py
View File

@@ -23,7 +23,6 @@ from transformers.testing_utils import (
is_pipeline_test,
nested_simplify,
require_av,
require_tf,
require_torch,
require_torch_or_tf,
require_vision,
@@ -124,8 +123,3 @@ class VideoClassificationPipelineTests(unittest.TestCase):
for output in outputs:
for element in output:
compare_pipeline_output_to_hub_spec(element, VideoClassificationOutputElement)
@require_tf
@unittest.skip
def test_small_model_tf(self):
pass

6
tests/pipelines/test_pipelines_visual_question_answering.py
View File

@@ -22,7 +22,6 @@ from transformers.testing_utils import (
is_pipeline_test,
is_torch_available,
nested_simplify,
require_tf,
require_torch,
require_torch_accelerator,
require_vision,
@@ -246,8 +245,3 @@ class VisualQuestionAnsweringPipelineTests(unittest.TestCase):
[{"score": ANY(float), "answer": ANY(str)}],
],
)
@require_tf
@unittest.skip(reason="Visual question answering not implemented in TF")
def test_small_model_tf(self):
pass

78
tests/pipelines/test_pipelines_zero_shot.py
View File

@@ -25,7 +25,6 @@ from transformers.testing_utils import (
is_pipeline_test,
is_torch_available,
nested_simplify,
require_tf,
require_torch,
slow,
)
@@ -243,26 +242,6 @@ class ZeroShotClassificationPipelineTests(unittest.TestCase):
},
)
@require_tf
def test_small_model_tf(self):
zero_shot_classifier = pipeline(
"zero-shot-classification",
model="sshleifer/tiny-distilbert-base-cased-distilled-squad",
framework="tf",
)
outputs = zero_shot_classifier(
"Who are you voting for in 2020?", candidate_labels=["politics", "public health", "science"]
)
self.assertEqual(
nested_simplify(outputs),
{
"sequence": "Who are you voting for in 2020?",
"labels": ["science", "public health", "politics"],
"scores": [0.333, 0.333, 0.333],
},
)
@slow
@require_torch
def test_large_model_pt(self):
@@ -319,60 +298,3 @@ class ZeroShotClassificationPipelineTests(unittest.TestCase):
"scores": [0.817, 0.713, 0.018, 0.018],
},
)
@slow
@require_tf
def test_large_model_tf(self):
zero_shot_classifier = pipeline(
"zero-shot-classification", model="FacebookAI/roberta-large-mnli", framework="tf"
)
outputs = zero_shot_classifier(
"Who are you voting for in 2020?", candidate_labels=["politics", "public health", "science"]
)
self.assertEqual(
nested_simplify(outputs),
{
"sequence": "Who are you voting for in 2020?",
"labels": ["politics", "public health", "science"],
"scores": [0.976, 0.015, 0.009],
},
)
outputs = zero_shot_classifier(
"The dominant sequence transduction models are based on complex recurrent or convolutional neural networks"
" in an encoder-decoder configuration. The best performing models also connect the encoder and decoder"
" through an attention mechanism. We propose a new simple network architecture, the Transformer, based"
" solely on attention mechanisms, dispensing with recurrence and convolutions entirely. Experiments on two"
" machine translation tasks show these models to be superior in quality while being more parallelizable"
" and requiring significantly less time to train. Our model achieves 28.4 BLEU on the WMT 2014"
" English-to-German translation task, improving over the existing best results, including ensembles by"
" over 2 BLEU. On the WMT 2014 English-to-French translation task, our model establishes a new"
" single-model state-of-the-art BLEU score of 41.8 after training for 3.5 days on eight GPUs, a small"
" fraction of the training costs of the best models from the literature. We show that the Transformer"
" generalizes well to other tasks by applying it successfully to English constituency parsing both with"
" large and limited training data.",
candidate_labels=["machine learning", "statistics", "translation", "vision"],
multi_label=True,
)
self.assertEqual(
nested_simplify(outputs),
{
"sequence": (
"The dominant sequence transduction models are based on complex recurrent or convolutional neural"
" networks in an encoder-decoder configuration. The best performing models also connect the"
" encoder and decoder through an attention mechanism. We propose a new simple network"
" architecture, the Transformer, based solely on attention mechanisms, dispensing with recurrence"
" and convolutions entirely. Experiments on two machine translation tasks show these models to be"
" superior in quality while being more parallelizable and requiring significantly less time to"
" train. Our model achieves 28.4 BLEU on the WMT 2014 English-to-German translation task,"
" improving over the existing best results, including ensembles by over 2 BLEU. On the WMT 2014"
" English-to-French translation task, our model establishes a new single-model state-of-the-art"
" BLEU score of 41.8 after training for 3.5 days on eight GPUs, a small fraction of the training"
" costs of the best models from the literature. We show that the Transformer generalizes well to"
" other tasks by applying it successfully to English constituency parsing both with large and"
" limited training data."
),
"labels": ["translation", "machine learning", "vision", "statistics"],
"scores": [0.817, 0.713, 0.018, 0.018],
},
)

83
tests/pipelines/test_pipelines_zero_shot_image_classification.py
View File

@@ -22,7 +22,6 @@ from transformers.testing_utils import (
compare_pipeline_output_to_hub_spec,
is_pipeline_test,
nested_simplify,
require_tf,
require_torch,
require_vision,
slow,
@@ -137,57 +136,6 @@ class ZeroShotImageClassificationPipelineTests(unittest.TestCase):
def test_small_model_pt_fp16(self):
self.test_small_model_pt(torch_dtype="float16")
@require_tf
def test_small_model_tf(self):
image_classifier = pipeline(
model="hf-internal-testing/tiny-random-clip-zero-shot-image-classification", framework="tf"
)
image = Image.open("./tests/fixtures/tests_samples/COCO/000000039769.png")
output = image_classifier(image, candidate_labels=["a", "b", "c"])
self.assertEqual(
nested_simplify(output),
[{"score": 0.333, "label": "a"}, {"score": 0.333, "label": "b"}, {"score": 0.333, "label": "c"}],
)
output = image_classifier([image] * 5, candidate_labels=["A", "B", "C"], batch_size=2)
self.assertEqual(
nested_simplify(output),
# Pipeline outputs are supposed to be deterministic and
# So we could in theory have real values "A", "B", "C" instead
# of ANY(str).
# However it seems that in this particular case, the floating
# scores are so close, we enter floating error approximation
# and the order is not guaranteed anymore with batching.
[
[
{"score": 0.333, "label": ANY(str)},
{"score": 0.333, "label": ANY(str)},
{"score": 0.333, "label": ANY(str)},
],
[
{"score": 0.333, "label": ANY(str)},
{"score": 0.333, "label": ANY(str)},
{"score": 0.333, "label": ANY(str)},
],
[
{"score": 0.333, "label": ANY(str)},
{"score": 0.333, "label": ANY(str)},
{"score": 0.333, "label": ANY(str)},
],
[
{"score": 0.333, "label": ANY(str)},
{"score": 0.333, "label": ANY(str)},
{"score": 0.333, "label": ANY(str)},
],
[
{"score": 0.333, "label": ANY(str)},
{"score": 0.333, "label": ANY(str)},
{"score": 0.333, "label": ANY(str)},
],
],
)
@slow
@require_torch
def test_large_model_pt(self):
@@ -221,37 +169,6 @@ class ZeroShotImageClassificationPipelineTests(unittest.TestCase):
* 5,
)
@slow
@require_tf
def test_large_model_tf(self):
image_classifier = pipeline(
task="zero-shot-image-classification", model="openai/clip-vit-base-patch32", framework="tf"
)
# This is an image of 2 cats with remotes and no planes
image = Image.open("./tests/fixtures/tests_samples/COCO/000000039769.png")
output = image_classifier(image, candidate_labels=["cat", "plane", "remote"])
self.assertEqual(
nested_simplify(output),
[
{"score": 0.511, "label": "remote"},
{"score": 0.485, "label": "cat"},
{"score": 0.004, "label": "plane"},
],
)
output = image_classifier([image] * 5, candidate_labels=["cat", "plane", "remote"], batch_size=2)
self.assertEqual(
nested_simplify(output),
[
[
{"score": 0.511, "label": "remote"},
{"score": 0.485, "label": "cat"},
{"score": 0.004, "label": "plane"},
],
]
* 5,
)
@slow
@require_torch
def test_siglip_model_pt(self):

11
tests/pipelines/test_pipelines_zero_shot_object_detection.py
View File

@@ -23,7 +23,6 @@ from transformers import (
from transformers.testing_utils import (
is_pipeline_test,
nested_simplify,
require_tf,
require_torch,
require_vision,
slow,
@@ -90,11 +89,6 @@ class ZeroShotObjectDetectionPipelineTests(unittest.TestCase):
],
)
@require_tf
@unittest.skip(reason="Zero Shot Object Detection not implemented in TF")
def test_small_model_tf(self):
pass
@require_torch
def test_small_model_pt(self):
object_detector = pipeline(
@@ -201,11 +195,6 @@ class ZeroShotObjectDetectionPipelineTests(unittest.TestCase):
],
)
@require_tf
@unittest.skip(reason="Zero Shot Object Detection not implemented in TF")
def test_large_model_tf(self):
pass
@require_torch
@slow
def test_threshold(self):