Update app.py

app.py

@@ -11,7 +11,7 @@ import os
 import time
 
 #Loading the model and the tokenizer
-#token_key = os.environ.get("HUGGING_FACE_HUB_TOKEN")
+token_key = os.environ.get("HUGGING_FACE_HUB_TOKEN")
 #model_name = "unilux/wav2vec-xls-r-Luxembourgish20-with-LM"
 #model_name = "unilux/wav2vec-xlsr-300m-Luxembourgish-with-LM"
 
@@ -26,7 +26,7 @@ model_name = "pgilles/whisper-large-v2-lb_cased_01"
 #processor = WhisperProcessor.from_pretrained(model_name)
 
 #p = pipeline("automatic-speech-recognition", model=model, tokenizer=tokenizer, feature_extractor=processor.feature_extractor, decoder=processor.decoder, use_auth_token=token_key)
-p = pipeline("automatic-speech-recognition", model=model_name, device=0)
+p = pipeline("automatic-speech-recognition", model=model_name, device=0, use_auth_token=token_key)
 
 #p = pipeline("automatic-speech-recognition", model=model_name, use_auth_token = token_key)
 #p = pipeline("automatic-speech-recognition", model=model_name, use_auth_token = True)
@@ -41,8 +41,8 @@ def load_data(input_file):
   sampling_rate = 16_000
   #read the file
   speech, sample_rate = librosa.load(input_file, sr=sampling_rate, mono=True)
-  speech = librosa.effects.trim(speech, top_db= 10)
-  return speech[0]
+  #speech = librosa.effects.trim(speech, top_db= 10)
+  return speech
     
 def asr_pipe(input_file):
   load_data(input_file)