Create script to transcript audios

transcribe.py

@@ -0,0 +1,26 @@
+import os
+import argparse
+
+def main(args):
+    audio_file = args.input
+    language = args.language
+    output_folder = "transcriptions"
+
+    # Transcribe audio file
+    model = "large-v2"
+    word_timestamps = True
+    fp16 = False
+    device = "cuda"
+    verbose = False
+    threads = 4
+    command = f'whisper --model {model} --output_dir {output_folder} --language {language} \
+--word_timestamps {word_timestamps} --fp16 {fp16} --device {device} --verbose {verbose} \
+--threads {threads} {audio_file}'
+    os.system(command)
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser(description='Transcribe audio files')
+    parser.add_argument('input', help='Input audio file')
+    parser.add_argument('language', help='Language of the audio file')
+    args = parser.parse_args()
+    main(args)