Update app.py

app.py

@@ -8,7 +8,6 @@ BATCH_SIZE = 8
 
 device = 0 if torch.cuda.is_available() else "cpu"
 
-# Initialize the pipeline
 pipe = pipeline(
     task="automatic-speech-recognition",
     model=MODEL_NAME,
@@ -16,20 +15,28 @@ pipe = pipeline(
     device=device,
 )
 
+
+# Copied from https://github.com/openai/whisper/blob/c09a7ae299c4c34c5839a76380ae407e7d785914/whisper/utils.py#L50
 def format_timestamp(seconds: float, always_include_hours: bool = False, decimal_marker: str = "."):
     if seconds is not None:
         milliseconds = round(seconds * 1000.0)
+
         hours = milliseconds // 3_600_000
         milliseconds -= hours * 3_600_000
+
         minutes = milliseconds // 60_000
         milliseconds -= minutes * 60_000
+
         seconds = milliseconds // 1_000
         milliseconds -= seconds * 1_000
+
         hours_marker = f"{hours:02d}:" if always_include_hours or hours > 0 else ""
         return f"{hours_marker}{minutes:02d}:{seconds:02d}{decimal_marker}{milliseconds:03d}"
     else:
+        # we have a malformed timestamp so just return it as is
         return seconds
 
+
 def transcribe(file, task, return_timestamps):
     outputs = pipe(file, batch_size=BATCH_SIZE, generate_kwargs={"task": task}, return_timestamps=return_timestamps)
     text = outputs["text"]
@@ -42,35 +49,53 @@ def transcribe(file, task, return_timestamps):
         text = "\n".join(str(feature) for feature in timestamps)
     return text
 
-# Use Blocks and modern Gradio components
-with gr.Blocks() as demo:
-    
-    with gr.TabbedInterface(["Transcribe Microphone", "Transcribe Audio File"]) as tabs:
-
-        with gr.TabItem("Transcribe Microphone"):
-            mic_audio = gr.Audio(source="microphone", type="filepath", label="Record Speech")
-            task = gr.Radio(["transcribe", "translate"], label="Task", value="transcribe")
-            return_timestamps = gr.Checkbox(label="Return timestamps")
-            mic_output = gr.Textbox(label="Transcription")
-            mic_button = gr.Button("Transcribe")
-            
-            mic_button.click(
-                fn=transcribe,
-                inputs=[mic_audio, task, return_timestamps],
-                outputs=mic_output,
-            )
-        
-        with gr.TabItem("Transcribe Audio File"):
-            file_audio = gr.Audio(source="upload", type="filepath", label="Upload Audio File")
-            task = gr.Radio(["transcribe", "translate"], label="Task", value="transcribe")
-            return_timestamps = gr.Checkbox(label="Return timestamps")
-            file_output = gr.Textbox(label="Transcription")
-            file_button = gr.Button("Transcribe")
-            
-            file_button.click(
-                fn=transcribe,
-                inputs=[file_audio, task, return_timestamps],
-                outputs=file_output,
-            )
-
-demo.launch(enable_queue=True)
+
+demo = gr.Blocks()
+
+mic_transcribe = gr.Interface(
+    fn=transcribe,
+    inputs=[
+        gr.Audio(source="microphone", type="filepath", optional=True),
+        gr.Radio(["transcribe", "translate"], label="Task", default="transcribe"),
+        gr.Checkbox(default=False, label="Return timestamps"),
+    ],
+    outputs="text",
+    layout="horizontal",
+    theme="huggingface",
+    title="Whisper Demo: Transcribe Audio",
+    description=(
+        "Transcribe long-form microphone or audio inputs with the click of a button! Demo uses the"
+        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.Audio(source="upload", optional=True, label="Audio file", type="filepath"),
+        gr.Radio(["transcribe", "translate"], label="Task", default="transcribe"),
+        gr.Checkbox(default=False, label="Return timestamps"),
+    ],
+    outputs="text",
+    layout="horizontal",
+    theme="huggingface",
+    title="Whisper Demo: Transcribe Audio",
+    description=(
+        "Transcribe long-form microphone or audio inputs with the click of a button! Demo uses the"
+        f" checkpoint [{MODEL_NAME}](https://huggingface.co/{MODEL_NAME}) and 🤗 Transformers to transcribe audio files"
+        " of arbitrary length."
+    ),
+    examples=[
+        ["./example.flac", "transcribe", False],
+        ["./example.flac", "transcribe", True],
+    ],
+    cache_examples=True,
+    allow_flagging="never",
+)
+
+with demo:
+    gr.TabbedInterface([mic_transcribe, file_transcribe], ["Transcribe Microphone", "Transcribe Audio File"])
+
+demo.launch(enable_queue=True)