def transcribe_audio_by_whisper(audio_path):

app.py

@@ -12,6 +12,8 @@ from openai import OpenAI
 import random
 import string
 
+from faster_whisper import WhisperModel
+
 
 # Initialize OpenAI API client
 OPEN_AI_KEY = os.getenv("OPEN_AI_KEY")
@@ -97,9 +99,26 @@ def download_youtube_audio(youtube_url):
 def transcribe_audio_by_whisper(audio_path):
     # Whisper模型的轉錄實現
     start_time = time.time()
+
+    model_size = "large-v3"
+    # Run on GPU with FP16
+    # model = WhisperModel(model_size, device="cuda", compute_type="float16")
+    # or run on GPU with INT8
+    # model = WhisperModel(model_size, device="cuda", compute_type="int8_float16")
+    # or run on CPU with INT8
+    model = WhisperModel(model_size, device="cpu", compute_type="int8")
+    segments, info = model.transcribe(audio_path, beam_size=5)
+    print("Detected language '%s' with probability %f" % (info.language, info.language_probability))
+
+    transcription = ""
+    for segment in segments:
+        print("[%.2fs -> %.2fs] %s" % (segment.start, segment.end, segment.text))
+        transcription += f"[{segment.start:.3f} -> {segment.end:.3f}] {segment.text}\n"
+    
     end_time = time.time()  # 函数结束执行的时间
     processing_time = int(end_time - start_time)
-    return "transcribe_audio_by_whisper", processing_time
+
+    return transcription, processing_time
 
 def transcribe_audio_by_open_ai(audio_path):
     # OpenAI語音識別的轉錄實現

requirements.txt

@@ -1,4 +1,5 @@
 yt-dlp
 gradio
 pydub
-openai >= 1.0.0
+openai >= 1.0.0
+faster-whisper