Update app.py

app.py

@@ -6,10 +6,10 @@ from faster_whisper import WhisperModel
 import ffmpeg
 
 # -------- Configuration --------
-MODEL_NAME = "small"  # choices: tiny, base, small, medium, large-v3
+MODEL_NAME = "small"  # tiny, base, small, medium, large-v3
 DEVICE = "cuda" if os.environ.get("USE_CUDA", "0") == "1" else "cpu"
 
-# Load model once
+# -------- Load Faster-Whisper --------
 print(f"🚀 Loading Faster-Whisper model: {MODEL_NAME} on {DEVICE}")
 model = WhisperModel(MODEL_NAME, device=DEVICE, compute_type="float16" if DEVICE == "cuda" else "int8")
 
@@ -27,9 +27,9 @@ def _format_timestamp(seconds: float) -> str:
 
 
 def transcribe(audio_file):
-    """Transcribe uploaded audio file and return text + SRT."""
+    """Transcribe uploaded audio and return text + SRT + file."""
     try:
-        # Convert to wav if needed (ensures consistency)
+        # Convert any format to wav for consistency
         with tempfile.NamedTemporaryFile(suffix=".wav", delete=False) as tmp_wav:
             (
                 ffmpeg
@@ -40,10 +40,9 @@ def transcribe(audio_file):
             )
             wav_path = tmp_wav.name
 
-        # Run transcription
+        # Transcribe
         segments, info = model.transcribe(wav_path, beam_size=5)
-        text_output = ""
-        srt_output = ""
+        text_output, srt_output = "", ""
 
         for i, segment in enumerate(segments, start=1):
             start = _format_timestamp(segment.start)
@@ -51,31 +50,46 @@ def transcribe(audio_file):
             srt_output += f"{i}\n{start} --> {end}\n{segment.text.strip()}\n\n"
             text_output += segment.text.strip() + " "
 
-        return text_output.strip(), srt_output
+        # Save SRT file
+        srt_path = Path(tempfile.mkstemp(suffix=".srt")[1])
+        with open(srt_path, "w", encoding="utf-8") as f:
+            f.write(srt_output)
+
+        return text_output.strip(), srt_output, srt_path
 
     except Exception as e:
-        return f"Error: {str(e)}", ""
+        return f"Error: {str(e)}", "", None
+
+
+def clear_outputs():
+    """Clear all UI fields."""
+    return None, "", "", None
 
 
 # -------- Gradio UI --------
 def build_ui():
-    with gr.Blocks(title="🎙️ Faster-Whisper Transcriber") as app:
-        gr.Markdown("# 🎧 Fast & Accurate Speech-to-Text using Faster-Whisper")
-        gr.Markdown("Upload or record an audio file to get instant transcription.")
+    with gr.Blocks(title="🎬 Subtitle Generator (Faster-Whisper)") as app:
+        gr.Markdown("# 🎧 Fast Subtitle Generator using Faster-Whisper")
+        gr.Markdown("Upload or record an audio file to generate `.srt` subtitles instantly.")
 
         with gr.Row():
-            audio_input = gr.Audio(sources=["upload", "microphone"], type="filepath", label="Audio Input")
+            audio_input = gr.Audio(sources=["upload", "microphone"], type="filepath", label="🎙️ Audio Input")
 
         with gr.Row():
-            text_output = gr.Textbox(label="Transcribed Text", lines=6)
-            srt_output = gr.Textbox(label="SRT Subtitle", lines=6)
+            text_output = gr.Textbox(label="📝 Transcribed Text", lines=6)
+            srt_output = gr.Textbox(label="📄 SRT Subtitle", lines=6)
+            srt_file = gr.File(label="⬇️ Download .srt File")
 
-        transcribe_btn = gr.Button("Transcribe")
+        with gr.Row():
+            transcribe_btn = gr.Button("🚀 Generate Subtitles")
+            clear_btn = gr.Button("🧹 Clear All")
 
-        transcribe_btn.click(fn=transcribe, inputs=audio_input, outputs=[text_output, srt_output])
+        # Button actions
+        transcribe_btn.click(fn=transcribe, inputs=audio_input, outputs=[text_output, srt_output, srt_file])
+        clear_btn.click(fn=clear_outputs, inputs=None, outputs=[audio_input, text_output, srt_output, srt_file])
 
         gr.Markdown("---")
-        gr.Markdown("Powered by **Faster-Whisper** ⚡")
+        gr.Markdown("⚡ Built with **Faster-Whisper** | 🎥 Ideal for Subtitle Generation")
 
     return app