added missing dependency in requirements.txt and upgraded to whisper-v3

handler.py

@@ -29,7 +29,7 @@ def get_audio_bytes(audio_path):
 
 class EndpointHandler:
     def __init__(self, path=""):
-        self.model = WhisperModel("large-v2", num_workers=30)
+        self.model = WhisperModel("large-v3", device="cpu")
 
     def __call__(self, data: dict[str, str]):
         inputs = data.pop("inputs")
@@ -50,7 +50,7 @@ class EndpointHandler:
             audio_bytes = io.BytesIO(audio_bytes_decoded)
 
         logging.info("Running inference...")
-        segments, info = self.model.transcribe(audio_bytes, language=language, task=task)
+        segments, info = self.model.transcribe(audio_bytes, language=language, task=task, )
 
         full_text = []
         for segment in segments:
@@ -86,14 +86,14 @@ if __name__ == '__main__':
     # Args.path = r"C:\Users\mbabu\AppData\Local\Temp\tmpsezkw2i5.mp3"
     # Args.path = "https://rr2---sn-4g5lzned.googlevideo.com/videoplayback?expire=1703474038&ei=Fp-IZeagJPaii9oPn4i3wAo&ip=195.146.4.71&id=o-ABVj7-vfJlewWZrzUMc466iPpKFkv2lNf6sHtRJ3F7s7&itag=22&source=youtube&requiressl=yes&xpc=EgVo2aDSNQ%3D%3D&mh=O5&mm=31%2C29&mn=sn-4g5lzned%2Csn-4g5ednsr&ms=au%2Crdu&mv=m&mvi=2&pl=25&initcwndbps=503750&spc=UWF9f8xjKc9m59EXQeaCQlNJLtGZUcI&vprv=1&svpuc=1&mime=video%2Fmp4&cnr=14&ratebypass=yes&dur=108.135&lmt=1701260382315374&mt=1703452143&fvip=1&fexp=24007246&c=ANDROID&txp=5308224&sparams=expire%2Cei%2Cip%2Cid%2Citag%2Csource%2Crequiressl%2Cxpc%2Cspc%2Cvprv%2Csvpuc%2Cmime%2Ccnr%2Cratebypass%2Cdur%2Clmt&sig=AJfQdSswRQIgP2osevilYOsP2YAHKqjlgQZVbT_UEzktBYVyOLUg4QgCIQDJdcRA_SOMBhwrMMVAkGkVyoho7rm99Y-io9xs1cVEsg%3D%3D&lsparams=mh%2Cmm%2Cmn%2Cms%2Cmv%2Cmvi%2Cpl%2Cinitcwndbps&lsig=AAO5W4owRgIhAJZ4g0mJIJzQ-5nvrNk5hdZQDzCfvifDuWfDXUu1tS0QAiEAuo-TgZtkwGUxLejKR7J_f2jU-aCV2pS8kx4Yl8zOQBc%3D&title=An%20alle%20Lehrkr%C3%A4fte%3A%20Finanzielle%20Bildung%20mit%20Daniel%20Jung%20%F0%9F%8E%93"
 
-    if is_cdn_link(Args.path):
-        test_inputs = Args.path
-    else:
-        audio = AudioSegment.from_mp3(Args.path)
-        buffer = io.BytesIO()
-        audio.export(buffer, format="mp3")
-        mp3_bytes = buffer.getvalue()
-        test_inputs = base64.b64encode(mp3_bytes)
+    # if is_cdn_link(Args.path):
+    #     test_inputs = Args.path
+    # else:
+    audio = AudioSegment.from_mp3(r"C:\Users\mbabu\AppData\Local\Temp\tmpsezkw2i5.mp3")
+    buffer = io.BytesIO()
+    audio.export(buffer, format="mp3")
+    mp3_bytes = buffer.getvalue()
+    test_inputs = base64.b64encode(mp3_bytes)
 
     sample_data = {
         "inputs": test_inputs,