Update app.py

comparison between implementations

app.py

@@ -78,6 +78,24 @@ def save_transcription(yt_url, transcription):
             json.dump({"url": yt_url, "transcription": transcription, "datetime": datetime.now().isoformat()}, f)
             f.write("\n")
 
+@spaces.GPU
+def yt_transcribe2(yt_url, task, max_filesize=75.0):
+    html_embed_str = _return_yt_html_embed(yt_url)
+
+    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()
+
+    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"]
+
+    return html_embed_str, text
+
+
 demo = gr.Blocks()
 
 yt_transcribe_interface = gr.Interface(
@@ -96,7 +114,23 @@ yt_transcribe_interface = gr.Interface(
     allow_flagging="never",
 )
 
+yt_transcribe = gr.Interface(
+    fn=yt_transcribe2,
+    inputs=[
+        gr.Textbox(lines=1, placeholder="Paste the URL to a YouTube video here", label="YouTube URL"),
+        gr.Radio(["transcribe", "translate"], label="Task", value="transcribe")
+    ],
+    outputs=["html", "text"],
+    title="Whisper Large V3: Transcribe YouTube",
+    description=(
+        "Transcribe long-form YouTube videos with the click of a button! Demo uses the checkpoint"
+        f" [{MODEL_NAME}](https://huggingface.co/{MODEL_NAME}) and 🤗 Transformers to transcribe video files of"
+        " arbitrary length."
+    ),
+    allow_flagging="never",
+)
+
 with demo:
-    gr.TabbedInterface([yt_transcribe_interface], ["YouTube"])
+    gr.TabbedInterface([yt_transcribe_interface, yt_transcribe], ["YouTube", "YouTube HF"])
 
 demo.queue().launch()