From fa39a9c45eb7b6a46b84e499ddc1299d5ff7ea45 Mon Sep 17 00:00:00 2001 From: Azalea Gui <22280294+hykilpikonna@users.noreply.github.com> Date: Wed, 11 Dec 2024 01:33:01 -0500 Subject: [PATCH] [F] Use large-v3 instead of distil-large-v3 for langauge support --- scripts/bin/transcribe | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/scripts/bin/transcribe b/scripts/bin/transcribe index c9c5473..4170ea3 100755 --- a/scripts/bin/transcribe +++ b/scripts/bin/transcribe @@ -9,7 +9,7 @@ from faster_whisper import WhisperModel, BatchedInferencePipeline # model_name = 'deepdml/faster-whisper-large-v3-turbo-ct2' -model_name = 'distil-large-v3' +model_name = 'large-v3' # model_name = 'medium.en' m = WhisperModel(model_name, device="cuda", compute_type="float16") model = BatchedInferencePipeline(model=m) @@ -54,7 +54,7 @@ def main(): parser = argparse.ArgumentParser(description="Transcribe audio from a video file and generate an SRT file.") # parser.add_argument("input_file", help="Path to the video file for transcription") parser.add_argument("input_file", nargs="+", help="Path to the video file for transcription") - parser.add_argument("-l", "--lang", default="en", help="Language code for transcription") + parser.add_argument("-l", "--lang", default=None, help="Language code for transcription (e.g. 'en')") args = parser.parse_args() for file in args.input_file: transcribe(Path(file), args.lang)