[Whisper] Fix slow test (#28407)

* [Whisper] Fix slow test

* update

* update

* update

* update

---------
Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>

.github/workflows/self-nightly-scheduled.yml

@@ -16,6 +16,7 @@ env:
   OMP_NUM_THREADS: 8
   MKL_NUM_THREADS: 8
   RUN_SLOW: yes
+  HF_HUB_READ_TOKEN: ${{ secrets.HF_HUB_READ_TOKEN }}
   SIGOPT_API_TOKEN: ${{ secrets.SIGOPT_API_TOKEN }}
   TF_FORCE_GPU_ALLOW_GROWTH: true
   RUN_PT_TF_CROSS_TESTS: 1

.github/workflows/self-past.yml

@@ -27,6 +27,7 @@ env:
   OMP_NUM_THREADS: 8
   MKL_NUM_THREADS: 8
   RUN_SLOW: yes
+  HF_HUB_READ_TOKEN: ${{ secrets.HF_HUB_READ_TOKEN }}
   SIGOPT_API_TOKEN: ${{ secrets.SIGOPT_API_TOKEN }}
   TF_FORCE_GPU_ALLOW_GROWTH: true
   RUN_PT_TF_CROSS_TESTS: 1

.github/workflows/self-push-amd.yml

@@ -15,6 +15,7 @@ env:
   PYTEST_TIMEOUT: 60
   TF_FORCE_GPU_ALLOW_GROWTH: true
   RUN_PT_TF_CROSS_TESTS: 1
+  HF_HUB_READ_TOKEN: ${{ secrets.HF_HUB_READ_TOKEN }}
 
 jobs:
   check_runner_status:

.github/workflows/self-scheduled-amd.yml

@@ -16,6 +16,7 @@ env:
   OMP_NUM_THREADS: 8
   MKL_NUM_THREADS: 8
   RUN_SLOW: yes
+  HF_HUB_READ_TOKEN: ${{ secrets.HF_HUB_READ_TOKEN }}
   SIGOPT_API_TOKEN: ${{ secrets.SIGOPT_API_TOKEN }}
 
 

.github/workflows/self-scheduled.yml

@@ -20,6 +20,9 @@ env:
   OMP_NUM_THREADS: 8
   MKL_NUM_THREADS: 8
   RUN_SLOW: yes
+  # For gated repositories, we still need to agree to share information on the Hub repo. page in order to get access.
+  # This token is created under the bot `hf-transformers-bot`.
+  HF_HUB_READ_TOKEN: ${{ secrets.HF_HUB_READ_TOKEN }}
   SIGOPT_API_TOKEN: ${{ secrets.SIGOPT_API_TOKEN }}
   TF_FORCE_GPU_ALLOW_GROWTH: true
   RUN_PT_TF_CROSS_TESTS: 1

tests/models/whisper/test_modeling_whisper.py

@@ -1651,8 +1651,10 @@ class WhisperModelIntegrationTests(unittest.TestCase):
         model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-large")
         model.to(torch_device)
 
-        ds = load_dataset("common_voice", "ja", split="test", streaming=True)
+        token = os.getenv("HF_HUB_READ_TOKEN", True)
+        ds = load_dataset("mozilla-foundation/common_voice_6_1", "ja", split="test", streaming=True, token=token)
         ds = ds.cast_column("audio", datasets.Audio(sampling_rate=16_000))
+
         input_speech = next(iter(ds))["audio"]["array"]
         input_features = processor.feature_extractor(raw_speech=input_speech, return_tensors="pt").input_features.to(
             torch_device