Remove trust_remote_code when loading Libri Dummy (#31748)

* [whisper integration] use parquet dataset for testing

* propagate to others

* more propagation

* last one

tests/pipelines/test_pipelines_audio_classification.py

@@ -71,9 +71,7 @@ class AudioClassificationPipelineTests(unittest.TestCase):
         import datasets
 
         # test with a local file
-        dataset = datasets.load_dataset(
-            "hf-internal-testing/librispeech_asr_dummy", "clean", split="validation", trust_remote_code=True
-        )
+        dataset = datasets.load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation")
         audio = dataset[0]["audio"]["array"]
         output = audio_classifier(audio)
         self.assertEqual(

tests/pipelines/test_pipelines_automatic_speech_recognition.py

@@ -294,9 +294,7 @@ class AutomaticSpeechRecognitionPipelineTests(unittest.TestCase):
         output = speech_recognizer(waveform)
         self.assertEqual(output, {"text": ""})
 
-        ds = load_dataset(
-            "hf-internal-testing/librispeech_asr_dummy", "clean", split="validation", trust_remote_code=True
-        ).sort("id")
+        ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation").sort("id")
         filename = ds[40]["file"]
         output = speech_recognizer(filename)
         self.assertEqual(output, {"text": "A MAN SAID TO THE UNIVERSE SIR I EXIST"})
@@ -313,9 +311,7 @@ class AutomaticSpeechRecognitionPipelineTests(unittest.TestCase):
         output = speech_recognizer(waveform)
         self.assertEqual(output, {"text": ""})
 
-        ds = load_dataset(
-            "hf-internal-testing/librispeech_asr_dummy", "clean", split="validation", trust_remote_code=True
-        ).sort("id")
+        ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation").sort("id")
         filename = ds[40]["file"]
         output = speech_recognizer(filename)
         self.assertEqual(output, {"text": "a man said to the universe sir i exist"})
@@ -545,9 +541,7 @@ class AutomaticSpeechRecognitionPipelineTests(unittest.TestCase):
             model="openai/whisper-tiny",
             framework="pt",
         )
-        ds = load_dataset(
-            "hf-internal-testing/librispeech_asr_dummy", "clean", split="validation", trust_remote_code=True
-        ).sort("id")
+        ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation").sort("id")
         filename = ds[40]["file"]
         output = speech_recognizer(filename)
         self.assertEqual(output, {"text": " A man said to the universe, Sir, I exist."})
@@ -722,9 +716,7 @@ class AutomaticSpeechRecognitionPipelineTests(unittest.TestCase):
     @slow
     @require_torch
     def test_whisper_timestamp_prediction(self):
-        ds = load_dataset(
-            "hf-internal-testing/librispeech_asr_dummy", "clean", split="validation", trust_remote_code=True
-        ).sort("id")
+        ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation").sort("id")
         array = np.concatenate(
             [ds[40]["audio"]["array"], ds[41]["audio"]["array"], ds[42]["audio"]["array"], ds[43]["audio"]["array"]]
         )
@@ -822,9 +814,7 @@ class AutomaticSpeechRecognitionPipelineTests(unittest.TestCase):
     @slow
     @require_torch
     def test_whisper_large_timestamp_prediction(self):
-        ds = load_dataset(
-            "hf-internal-testing/librispeech_asr_dummy", "clean", split="validation", trust_remote_code=True
-        ).sort("id")
+        ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation").sort("id")
         array = np.concatenate(
             [ds[40]["audio"]["array"], ds[41]["audio"]["array"], ds[42]["audio"]["array"], ds[43]["audio"]["array"]]
         )
@@ -918,9 +908,7 @@ class AutomaticSpeechRecognitionPipelineTests(unittest.TestCase):
             chunk_length_s=3,
             return_timestamps="word",
         )
-        data = load_dataset(
-            "hf-internal-testing/librispeech_asr_dummy", "clean", split="validation", trust_remote_code=True
-        )
+        data = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation")
         sample = data[0]["audio"]
 
         # not the same output as test_simple_whisper_asr because of chunking
@@ -963,9 +951,7 @@ class AutomaticSpeechRecognitionPipelineTests(unittest.TestCase):
             model="openai/whisper-large-v3",
             return_timestamps="word",
         )
-        data = load_dataset(
-            "hf-internal-testing/librispeech_asr_dummy", "clean", split="validation", trust_remote_code=True
-        )
+        data = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation")
         sample = data[0]["audio"]
 
         # not the same output as test_simple_whisper_asr because of chunking
@@ -1010,9 +996,7 @@ class AutomaticSpeechRecognitionPipelineTests(unittest.TestCase):
             framework="pt",
         )
 
-        ds = load_dataset(
-            "hf-internal-testing/librispeech_asr_dummy", "clean", split="validation", trust_remote_code=True
-        ).sort("id")
+        ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation").sort("id")
         filename = ds[40]["file"]
         output = speech_recognizer(filename)
         self.assertEqual(output, {"text": 'Ein Mann sagte zum Universum : " Sir, ich existiert! "'})
@@ -1030,9 +1014,7 @@ class AutomaticSpeechRecognitionPipelineTests(unittest.TestCase):
         output = asr(waveform)
         self.assertEqual(output, {"text": ""})
 
-        ds = load_dataset(
-            "hf-internal-testing/librispeech_asr_dummy", "clean", split="validation", trust_remote_code=True
-        ).sort("id")
+        ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation").sort("id")
         filename = ds[40]["file"]
         output = asr(filename)
         self.assertEqual(output, {"text": "A MAN SAID TO THE UNIVERSE SIR I EXIST"})
@@ -1058,9 +1040,7 @@ class AutomaticSpeechRecognitionPipelineTests(unittest.TestCase):
         output = asr(waveform)
         self.assertEqual(output, {"text": "(Applausi)"})
 
-        ds = load_dataset(
-            "hf-internal-testing/librispeech_asr_dummy", "clean", split="validation", trust_remote_code=True
-        ).sort("id")
+        ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation").sort("id")
         filename = ds[40]["file"]
         output = asr(filename)
         self.assertEqual(output, {"text": "Un uomo disse all'universo: \"Signore, io esisto."})
@@ -1080,9 +1060,7 @@ class AutomaticSpeechRecognitionPipelineTests(unittest.TestCase):
             model="openai/whisper-tiny.en",
             framework="pt",
         )
-        ds = load_dataset(
-            "hf-internal-testing/librispeech_asr_dummy", "clean", split="validation", trust_remote_code=True
-        )
+        ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation")
         filename = ds[0]["file"]
         output = speech_recognizer(filename)
         self.assertEqual(
@@ -1151,9 +1129,7 @@ class AutomaticSpeechRecognitionPipelineTests(unittest.TestCase):
             model="openai/whisper-large",
             framework="pt",
         )
-        ds = load_dataset(
-            "hf-internal-testing/librispeech_asr_dummy", "clean", split="validation", trust_remote_code=True
-        ).sort("id")
+        ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation").sort("id")
         filename = ds[40]["file"]
         output = speech_recognizer(filename)
         self.assertEqual(output, {"text": " A man said to the universe, Sir, I exist."})
@@ -1188,9 +1164,7 @@ class AutomaticSpeechRecognitionPipelineTests(unittest.TestCase):
             model="openai/whisper-tiny.en",
             framework="pt",
         )
-        ds = load_dataset(
-            "hf-internal-testing/librispeech_asr_dummy", "clean", split="validation", trust_remote_code=True
-        )
+        ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation")
         filename = ds[0]["file"]
 
         # 1. English-only model compatible with no language argument
@@ -1323,9 +1297,7 @@ class AutomaticSpeechRecognitionPipelineTests(unittest.TestCase):
             framework="pt",
         )
 
-        ds = load_dataset(
-            "hf-internal-testing/librispeech_asr_dummy", "clean", split="validation", trust_remote_code=True
-        ).sort("id")
+        ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation").sort("id")
         filename = ds[40]["file"]
         output = speech_recognizer(filename)
         self.assertEqual(output, {"text": "A man said to the universe: “Sir, I exist."})
@@ -1341,9 +1313,7 @@ class AutomaticSpeechRecognitionPipelineTests(unittest.TestCase):
             framework="pt",
         )
 
-        ds = load_dataset(
-            "hf-internal-testing/librispeech_asr_dummy", "clean", split="validation", trust_remote_code=True
-        ).sort("id")
+        ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation").sort("id")
         filename = ds[40]["file"]
         output = speech_recognizer(filename)
         self.assertEqual(output, {"text": "Ein Mann sagte zu dem Universum, Sir, ich bin da."})
@@ -1360,9 +1330,7 @@ class AutomaticSpeechRecognitionPipelineTests(unittest.TestCase):
             framework="pt",
         )
 
-        ds = load_dataset(
-            "hf-internal-testing/librispeech_asr_dummy", "clean", split="validation", trust_remote_code=True
-        ).sort("id")
+        ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation").sort("id")
         filename = ds[40]["file"]
 
         output = speech_recognizer(filename)
@@ -1379,9 +1347,7 @@ class AutomaticSpeechRecognitionPipelineTests(unittest.TestCase):
             framework="pt",
         )
 
-        dataset = load_dataset(
-            "hf-internal-testing/librispeech_asr_dummy", "clean", split="validation", trust_remote_code=True
-        )
+        dataset = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation")
         sample = dataset[0]["audio"]
 
         output = speech_recognizer(sample)
@@ -1398,9 +1364,7 @@ class AutomaticSpeechRecognitionPipelineTests(unittest.TestCase):
             chunk_length_s=10.0,
         )
 
-        ds = load_dataset(
-            "hf-internal-testing/librispeech_asr_dummy", "clean", split="validation", trust_remote_code=True
-        ).sort("id")
+        ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation").sort("id")
         audio = ds[40]["audio"]["array"]
 
         n_repeats = 2
@@ -1416,9 +1380,7 @@ class AutomaticSpeechRecognitionPipelineTests(unittest.TestCase):
             model="hf-internal-testing/tiny-random-wav2vec2",
         )
 
-        ds = load_dataset(
-            "hf-internal-testing/librispeech_asr_dummy", "clean", split="validation", trust_remote_code=True
-        ).sort("id")
+        ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation").sort("id")
         # Take short audio to keep the test readable
         audio = ds[40]["audio"]["array"][:800]
 
@@ -1462,9 +1424,7 @@ class AutomaticSpeechRecognitionPipelineTests(unittest.TestCase):
             chunk_length_s=10.0,
         )
 
-        ds = load_dataset(
-            "hf-internal-testing/librispeech_asr_dummy", "clean", split="validation", trust_remote_code=True
-        ).sort("id")
+        ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation").sort("id")
         audio = ds[40]["audio"]["array"]
 
         n_repeats = 2
@@ -1492,9 +1452,7 @@ class AutomaticSpeechRecognitionPipelineTests(unittest.TestCase):
         )
         self.assertEqual(speech_recognizer.type, "ctc_with_lm")
 
-        ds = load_dataset(
-            "hf-internal-testing/librispeech_asr_dummy", "clean", split="validation", trust_remote_code=True
-        ).sort("id")
+        ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation").sort("id")
         audio = ds[40]["audio"]["array"]
 
         n_repeats = 2
@@ -1522,9 +1480,7 @@ class AutomaticSpeechRecognitionPipelineTests(unittest.TestCase):
         )
         self.assertEqual(speech_recognizer.type, "ctc_with_lm")
 
-        ds = load_dataset(
-            "hf-internal-testing/librispeech_asr_dummy", "clean", split="validation", trust_remote_code=True
-        ).sort("id")
+        ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation").sort("id")
         audio = ds[40]["audio"]["array"]
 
         n_repeats = 2
@@ -1608,9 +1564,7 @@ class AutomaticSpeechRecognitionPipelineTests(unittest.TestCase):
             device=torch_device,
         )
 
-        dataset = load_dataset(
-            "hf-internal-testing/librispeech_asr_dummy", "clean", split="validation", trust_remote_code=True
-        )
+        dataset = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation")
         sample = dataset[0]["audio"]
 
         result = pipe(sample, generate_kwargs={"tgt_lang": "eng"})
@@ -1633,9 +1587,7 @@ class AutomaticSpeechRecognitionPipelineTests(unittest.TestCase):
             chunk_length_s=10.0,
         )
 
-        ds = load_dataset(
-            "hf-internal-testing/librispeech_asr_dummy", "clean", split="validation", trust_remote_code=True
-        ).sort("id")
+        ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation").sort("id")
         audio = ds[40]["audio"]["array"]
 
         n_repeats = 10
@@ -1747,9 +1699,7 @@ class AutomaticSpeechRecognitionPipelineTests(unittest.TestCase):
             model="patrickvonplaten/wav2vec2-base-100h-with-lm",
             chunk_length_s=10.0,
         )
-        ds = load_dataset(
-            "hf-internal-testing/librispeech_asr_dummy", "clean", split="validation", trust_remote_code=True
-        ).sort("id")
+        ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation").sort("id")
         audio = ds[40]["audio"]["array"]
 
         n_repeats = 10

tests/pipelines/test_pipelines_common.py

@@ -840,9 +840,7 @@ class CustomPipelineTest(unittest.TestCase):
     def test_chunk_pipeline_batching_single_file(self):
         # Make sure we have cached the pipeline.
         pipe = pipeline(model="hf-internal-testing/tiny-random-Wav2Vec2ForCTC")
-        ds = datasets.load_dataset(
-            "hf-internal-testing/librispeech_asr_dummy", "clean", split="validation", trust_remote_code=True
-        ).sort("id")
+        ds = datasets.load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation").sort("id")
         audio = ds[40]["audio"]["array"]
 
         pipe = pipeline(model="hf-internal-testing/tiny-random-Wav2Vec2ForCTC")