Enhance batch_confirm_hf_optimized.py to ensure torchcodec is installed before loading the dataset, and update requirements_optimized.txt to include torchcodec. Modify run_optimized_192cores_no_root.sh to install additional audio dependencies and test audio imports.

vosk/test_files/test_audio_deps.py

@@ -0,0 +1,42 @@
+#!/usr/bin/env python3
+"""
+Test script for audio dependencies
+"""
+
+def test_audio_dependencies():
+    """Test if all audio dependencies are installed"""
+    try:
+        import torchcodec
+        print("torchcodec: OK")
+    except ImportError as e:
+        print(f"torchcodec: FAILED - {e}")
+        return False
+    
+    try:
+        import datasets
+        print("datasets: OK")
+    except ImportError as e:
+        print(f"datasets: FAILED - {e}")
+        return False
+    
+    try:
+        import soundfile
+        print("soundfile: OK")
+    except ImportError as e:
+        print(f"soundfile: FAILED - {e}")
+        return False
+    
+    try:
+        import librosa
+        print("librosa: OK")
+    except ImportError as e:
+        print(f"librosa: FAILED - {e}")
+        return False
+    
+    print("All audio dependencies installed successfully")
+    return True
+
+if __name__ == "__main__":
+    success = test_audio_dependencies()
+    if not success:
+        exit(1)