From add784a02f64ae9afcbbaca31a2c4df037ee0922 Mon Sep 17 00:00:00 2001
From: Shadow <124238783+ShadowLoveElysia@users.noreply.github.com>
Date: Mon, 5 Feb 2024 18:09:57 +0800
Subject: [PATCH] Add files via upload

python WhisperASR.py -i <input> -o <out_put> -f <file_name.list> -l <language>
---
 tools/damo_asr/WhisperASR.py | 42 ++++++++++++++++++++++++++++++++++++
 1 file changed, 42 insertions(+)
 create mode 100644 tools/damo_asr/WhisperASR.py

diff --git a/tools/damo_asr/WhisperASR.py b/tools/damo_asr/WhisperASR.py
new file mode 100644
index 0000000..3b0a946
--- /dev/null
+++ b/tools/damo_asr/WhisperASR.py
@@ -0,0 +1,42 @@
+import os
+import argparse
+import os
+os.environ["KMP_DUPLICATE_LIB_OK"]="TRUE"
+from glob import glob
+from faster_whisper import WhisperModel
+
+def main(input_folder, output_folder, output_filename, language):
+    model = WhisperModel("large-v3", device="cuda", compute_type="float16")
+
+    output_file = os.path.join(output_folder, output_filename)
+    if not os.path.exists(output_folder):
+        os.makedirs(output_folder)
+
+    with open(output_file, 'w', encoding='utf-8') as f:
+        for file in glob(os.path.join(input_folder, '**/*.wav'), recursive=True):
+            segments, _ = model.transcribe(file, beam_size=10, vad_filter=True,
+                                           vad_parameters=dict(min_silence_duration_ms=700), language=language)
+            segments = list(segments)
+
+            filename = os.path.basename(file).replace('.wav', '')
+            directory = os.path.dirname(file)
+
+            result_line = f"{file}|{language.upper()}|{segments[0].text}\n"
+            f.write(result_line)
+
+if __name__ == '__main__':
+    parser = argparse.ArgumentParser()
+    parser.add_argument("-i", "--input_folder", type=str, required=True,
+                        help="Path to the folder containing WAV files.")
+    parser.add_argument("-o", "--output_folder", type=str, required=True, help="Output folder to store transcriptions.")
+    parser.add_argument("-f", "--output_filename", type=str, default="transcriptions.txt", help="Name of the output text file.")
+    parser.add_argument("-l", "--language", type=str, default='zh', choices=['zh', 'en', ...],
+                        help="Language of the audio files.")
+
+    cmd = parser.parse_args()
+
+    input_folder = cmd.input_folder
+    output_folder = cmd.output_folder
+    output_filename = cmd.output_filename
+    language = cmd.language
+    main(input_folder, output_folder, output_filename, language)
\ No newline at end of file