Feature request: added huggingface functionality checking for token

src/main.py

@@ -1,3 +1,5 @@
+import os
+
 from src.transcript_generator import TranscriptGenerator
 
 
@@ -7,6 +9,11 @@ def main():
         'https://www.youtube.com/watch?v=BVVo1BaL6JA'
     ]
 
+    hf_token = os.environ.get('HF_TOKEN')
+    if not hf_token:
+        hf_token = input("Please enter your HF_TOKEN: ")
+    os.environ['HF_TOKEN'] = hf_token
+
     transcript_generator = TranscriptGenerator()
     transcript_generator.generate(input_sources)
 

src/transcriber.py

@@ -1,6 +1,7 @@
 import gc
 import json
 import logging
+import os
 
 import torch
 import whisperx
@@ -30,6 +31,7 @@ class Transcriber:
             logger.exception("Failed to initialize models: %s", e)
             raise
         self.transcript_output = None
+        self.hf_token = os.environ.get('HF_TOKEN')
 
     def transcribe(self, wav_filename: str, batch_size: int = 16) -> dict:
         """
@@ -53,7 +55,7 @@ class Transcriber:
             self._clear_memory()
 
             # Run Diarization
-            diarize_model = whisperx.DiarizationPipeline(use_auth_token='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',
+            diarize_model = whisperx.DiarizationPipeline(use_auth_token=self.hf_token,
                                                          device=self.device)
             diarize_segments = diarize_model(audio)
             result = whisperx.assign_word_speakers(diarize_segments, result)