Update app.py

app.py

@@ -68,12 +68,12 @@ DEFAULT_TARGET_LANGUAGE = "Bengali"
 @spaces.GPU
 def run_asr_ctc(input_audio: str, target_language: str) -> str:
     # preprocess_audio(input_audio)
-    input_audio, orig_freq = torchaudio.load(input_audio)
-    input_audio = torchaudio.functional.resample(input_audio, orig_freq=orig_freq, new_freq=16000)
+    # input_audio, orig_freq = torchaudio.load(input_audio)
+    # input_audio = torchaudio.functional.resample(input_audio, orig_freq=orig_freq, new_freq=16000)
     lang_id = LANGUAGE_NAME_TO_CODE[target_language]
 
     model.cur_decoder = "ctc"
-    ctc_text = model.transcribe(['sample_audio_infer_ready.wav'], batch_size=1,logprobs=False, language_id=lang_id)[0]
+    ctc_text = model.transcribe([input_audio], batch_size=1, logprobs=False, language_id=lang_id)[0]
 
     return ctc_text
 
@@ -85,7 +85,7 @@ def run_asr_rnnt(input_audio: str, target_language: str) -> str:
     lang_id = LANGUAGE_NAME_TO_CODE[target_language]
 
     model.cur_decoder = "rnnt"
-    ctc_text = model.transcribe(['sample_audio_infer_ready.wav'], batch_size=1,logprobs=False, language_id=lang_id)[0]
+    ctc_text = model.transcribe([input_audio], batch_size=1,logprobs=False, language_id=lang_id)[0]
 
     return ctc_text