S-Fry
/

large

Automatic Speech Recognition

hf-asr-leaderboard

Inference Endpoints

Model card Files Files and versions Community

S-Fry commited on Apr 27, 2023

Commit

109fb39

•

1 Parent(s): 715bca3

Update handler.py

Files changed (1) hide show

handler.py +1 -1

handler.py CHANGED Viewed

@@ -19,7 +19,7 @@ class EndpointHandler():
         #ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation")
         #sample = ds[0]["audio"]
         inputs = data.pop("inputs", data)
-        audio_nparray = ffmpeg_read(inputs, sample_rate=16000)
         audio_tensor = torch.from_numpy(audio_nparray)
         prediction = pipe(audio_nparray, return_timestamps=True)

         #ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation")
         #sample = ds[0]["audio"]
         inputs = data.pop("inputs", data)
+        audio_nparray = ffmpeg_read(inputs, 16000)
         audio_tensor = torch.from_numpy(audio_nparray)
         prediction = pipe(audio_nparray, return_timestamps=True)