Update app.py

text = pipe(inputs, batch_size=BATCH_SIZE, generate_kwargs={"task": task}, return_timestamps=True)

app.py

@@ -1,10 +1,17 @@
 import torch
-# from PIL import Image
+
 import gradio as gr
-import pytube as pt
+import yt_dlp as youtube_dl
 from transformers import pipeline
+from transformers.pipelines.audio_utils import ffmpeg_read
+
+import tempfile
+import os
 
 MODEL_NAME = "openai/whisper-large-v3"
+BATCH_SIZE = 8
+# FILE_LIMIT_MB = 1000
+# YT_LENGTH_LIMIT_S = 3600  # limit to 1 hour YouTube files
 
 device = 0 if torch.cuda.is_available() else "cpu"
 
@@ -16,32 +23,12 @@ pipe = pipeline(
 )
 
 
-all_special_ids = pipe.tokenizer.all_special_ids
-transcribe_token_id = all_special_ids[-5]
-translate_token_id = all_special_ids[-6]
-
-
-def transcribe(microphone, file_upload, task):
-    warn_output = ""
-    if (microphone is not None) and (file_upload is not None):
-        warn_output = (
-            "警告：您已经上传了一个音频文件并使用了麦克录制。"
-            "录制文件将被使用上传的音频将被丢弃。\n"
-        )
-
-    elif (microphone is None) and (file_upload is None):
-        return "错误: 您必须使用麦克风录制或上传音频文件"
+def transcribe(inputs, task):
+    if inputs is None:
+        raise gr.Error("No audio file submitted! Please upload or record an audio file before submitting your request.")
 
-    file = microphone if microphone is not None else file_upload
-
-    pipe.model.config.forced_decoder_ids = [[2, transcribe_token_id if task=="transcribe" else translate_token_id]]
-
-    # text = pipe(file, return_timestamps=True)["text"]
-    text = pipe(file, return_timestamps=True)
-    #trans to SRT
-    text= convert_to_srt(text)
-    
-    return warn_output + text
+    text = pipe(inputs, batch_size=BATCH_SIZE, generate_kwargs={"task": task}, return_timestamps=True)["text"]
+    return  text
 
 
 def _return_yt_html_embed(yt_url):
@@ -52,22 +39,61 @@ def _return_yt_html_embed(yt_url):
     )
     return HTML_str
 
+def download_yt_audio(yt_url, filename):
+    info_loader = youtube_dl.YoutubeDL()
+    
+    try:
+        info = info_loader.extract_info(yt_url, download=False)
+    except youtube_dl.utils.DownloadError as err:
+        raise gr.Error(str(err))
+    
+    file_length = info["duration_string"]
+    file_h_m_s = file_length.split(":")
+    file_h_m_s = [int(sub_length) for sub_length in file_h_m_s]
+    
+    if len(file_h_m_s) == 1:
+        file_h_m_s.insert(0, 0)
+    if len(file_h_m_s) == 2:
+        file_h_m_s.insert(0, 0)
+    file_length_s = file_h_m_s[0] * 3600 + file_h_m_s[1] * 60 + file_h_m_s[2]
+    
+    if file_length_s > YT_LENGTH_LIMIT_S:
+        yt_length_limit_hms = time.strftime("%HH:%MM:%SS", time.gmtime(YT_LENGTH_LIMIT_S))
+        file_length_hms = time.strftime("%HH:%MM:%SS", time.gmtime(file_length_s))
+        raise gr.Error(f"Maximum YouTube length is {yt_length_limit_hms}, got {file_length_hms} YouTube video.")
+    
+    ydl_opts = {"outtmpl": filename, "format": "worstvideo[ext=mp4]+bestaudio[ext=m4a]/best[ext=mp4]/best"}
+    
+    with youtube_dl.YoutubeDL(ydl_opts) as ydl:
+        try:
+            ydl.download([yt_url])
+        except youtube_dl.utils.ExtractorError as err:
+            raise gr.Error(str(err))
+
 
 def yt_transcribe(yt_url, task):
-    yt = pt.YouTube(yt_url)
     html_embed_str = _return_yt_html_embed(yt_url)
-    stream = yt.streams.filter(only_audio=True)[0]
-    stream.download(filename="audio.mp3")
 
-    pipe.model.config.forced_decoder_ids = [[2, transcribe_token_id if task=="transcribe" else translate_token_id]]
+    with tempfile.TemporaryDirectory() as tmpdirname:
+        filepath = os.path.join(tmpdirname, "video.mp4")
+        download_yt_audio(yt_url, filepath)
+        with open(filepath, "rb") as f:
+            inputs = f.read()
 
-    text = pipe("audio.mp3",return_timestamps=True)
-    # text = pipe("audio.mp3",return_timestamps=True)["text"]
+    inputs = ffmpeg_read(inputs, pipe.feature_extractor.sampling_rate)
+    inputs = {"array": inputs, "sampling_rate": pipe.feature_extractor.sampling_rate}
+
+    
+    text = pipe(inputs, batch_size=BATCH_SIZE, generate_kwargs={"task": task}, return_timestamps=True)
+
+    # text = pipe(inputs, batch_size=BATCH_SIZE, generate_kwargs={"task": task}, return_timestamps=True)["text"]
+    # text = pipe("audio.mp3",return_timestamps=True)
     #trans to SRT
     text= convert_to_srt(text)
-
     return html_embed_str, text
 
+    
+# SRT prepare
 # Assuming srt format is a sequence of subtitles with index, time range and text
 def convert_to_srt(input):
   output = ""
@@ -94,23 +120,41 @@ def format_time(seconds):
   milliseconds = int((seconds % 1) * 1000)
   return f"{hours:02}:{minutes:02}:{seconds:02},{milliseconds:03}"
 
+
 demo = gr.Blocks()
 
 mf_transcribe = gr.Interface(
     fn=transcribe,
     inputs=[
         gr.inputs.Audio(source="microphone", type="filepath", optional=True),
-        gr.inputs.Audio(source="upload", type="filepath", optional=True),
         gr.inputs.Radio(["transcribe", "translate"], label="Task", default="transcribe"),
     ],
     outputs="text",
     layout="horizontal",
     theme="huggingface",
-    title="Audio-to-Text-SRT 自动生成字幕",
+    title="Whisper Large V3: Transcribe Audio",
+    description=(
+        "Transcribe long-form microphone or audio inputs with the click of a button! Demo uses the OpenAI Whisper"
+        f" checkpoint [{MODEL_NAME}](https://huggingface.co/{MODEL_NAME}) and 🤗 Transformers to transcribe audio files"
+        " of arbitrary length."
+    ),
+    allow_flagging="never",
+)
+
+file_transcribe = gr.Interface(
+    fn=transcribe,
+    inputs=[
+        gr.inputs.Audio(source="upload", type="filepath", optional=True, label="Audio file"),
+        gr.inputs.Radio(["transcribe", "translate"], label="Task", default="transcribe"),
+    ],
+    outputs="text",
+    layout="horizontal",
+    theme="huggingface",
+    title="Whisper Large V3: Transcribe Audio",
     description=(
-        "直接在网页录音或上传音频文件，加入Youtube连接，轻松转换为文字和字幕格式! 本演示采用"
-        f" 模型 [{MODEL_NAME}](https://huggingface.co/{MODEL_NAME}) 和 🤗 Transformers 转换任意长度的"
-        "音视频文件!使用GPU转换效率会大幅提高，大约每小时 $0.6 约相当于人民币 5 元。如果您有较长内容,需要更快的转换速度，请私信作者微信 1259388，并备注“语音转文字”。"
+        "Transcribe long-form microphone or audio inputs with the click of a button! Demo uses the OpenAI Whisper"
+        f" checkpoint [{MODEL_NAME}](https://huggingface.co/{MODEL_NAME}) and 🤗 Transformers to transcribe audio files"
+        " of arbitrary length."
     ),
     allow_flagging="never",
 )
@@ -119,32 +163,21 @@ yt_transcribe = gr.Interface(
     fn=yt_transcribe,
     inputs=[
         gr.inputs.Textbox(lines=1, placeholder="Paste the URL to a YouTube video here", label="YouTube URL"),
-        gr.inputs.Radio(["转译", "翻译"], label="Task", default="transcribe")
+        gr.inputs.Radio(["transcribe", "translate"], label="Task", default="transcribe")
     ],
     outputs=["html", "text"],
     layout="horizontal",
     theme="huggingface",
-    title="Audio-to-Text-SRT 自动生成字幕",
+    title="Whisper Large V3: Transcribe YouTube",
     description=(
-        "直接在网页录音或上传音频文件，加入Youtube连接，轻松转换为文字和字幕格式! 本演示采用"
-        f" 模型 [{MODEL_NAME}](https://huggingface.co/{MODEL_NAME}) 和 🤗 Transformers 转换任意长度的"
-        "音视频文件!使用GPU转换效率会大幅提高，大约每小时 $0.6 约相当于人民币 5 元。如果您有较长内容,需要更快的转换速度，请私信作者微信 1259388，并备注“语音转文字”。"
+        "Transcribe long-form YouTube videos with the click of a button! Demo uses the OpenAI Whisper checkpoint"
+        f" [{MODEL_NAME}](https://huggingface.co/{MODEL_NAME}) and 🤗 Transformers to transcribe video files of"
+        " arbitrary length."
     ),
     allow_flagging="never",
 )
 
-#  # Load the images
-# image1 = Image("wechatqrcode.jpg")
-# image2 = Image("paypalqrcode.png")
-    
-#     # Define a function that returns the images and captions
-# def display_images():
-#     return image1, "WeChat Pay", image2, "PayPal"
-    
 with demo:
-    gr.TabbedInterface([mf_transcribe, yt_transcribe], ["转译音频成文字", "YouTube转字幕"])
-       
-    # Create a gradio interface with no inputs and four outputs
-    # gr.Interface(display_images, [], [gr.outputs.Image(), gr.outputs.Textbox(), gr.outputs.Image(), gr.outputs.Textbox()], layout="horizontal").launch()
+    gr.TabbedInterface([mf_transcribe, file_transcribe, yt_transcribe], ["Microphone", "Audio file", "YouTube"])
 
 demo.launch(enable_queue=True)