Rename tools/extract/asr.py to tools/extract/asr_whisperx.py

tools/extract/{asr.py → asr_whisperx.py}

@@ -1,4 +1,3 @@
-import inspect
 from pathlib import Path
 
 import torch
@@ -22,22 +21,12 @@ class ASRProcessor:
 
     def __init__(self, model_name="large-v2", compute_type="float16"):
         self.model_name = model_name
-        # Check if whisperx.load_model accepts compute_type parameter
-
-        if "compute_type" in inspect.signature(whisperx.load_model).parameters:
-            self.model = whisperx.load_model(
-                model_name, device, compute_type=compute_type
-            )
-        else:
-            self.model = whisperx.load_model(model_name, device)
+        self.model = whisperx.load_model(model_name, device, compute_type=compute_type)
 
     def get_asr(self, audio_file, return_duration=True):
         assert Path(audio_file).exists(), f"File {audio_file} does not exist"
         audio = whisperx.load_audio(audio_file)
-        if "batch_size" in inspect.signature(self.model.transcribe).parameters:
-            result = self.model.transcribe(audio, batch_size=1)
-        else:
-            result = self.model.transcribe(audio)
+        result = self.model.transcribe(audio, batch_size=1)
         language = result["language"]
         duration = audio.shape[0] / SAMPLE_RATE