Update app.py

app.py

@@ -1,9 +1,94 @@
-from TTS.api import TTS
+import gradio as gr
 import torch
+import torchaudio
+import tempfile
+import os
+from TTS.api import TTS
 
+# ---------------------------
+# Device setup (CPU only)
+# ---------------------------
 device = "cpu"
 
+print("Loading XTTS model...")
 tts = TTS(
     model_name="tts_models/multilingual/multi-dataset/xtts_v2",
     progress_bar=False
-).to(device)
+).to(device)
+print("Model loaded!")
+
+# ---------------------------
+# Voice cloning function
+# ---------------------------
+def clone_voice(audio_file, text, lang):
+    try:
+        if audio_file is None:
+            return None, "❌ Please upload audio"
+
+        if text.strip() == "":
+            return None, "❌ Please enter text"
+
+        # CPU safety limit
+        if len(text) > 200:
+            return None, "❌ Text too long (max 200 chars for CPU)"
+
+        # Load audio
+        waveform, sr = torchaudio.load(audio_file)
+
+        # Convert to mono
+        if waveform.shape[0] > 1:
+            waveform = waveform.mean(dim=0, keepdim=True)
+
+        # Save temp speaker audio
+        with tempfile.NamedTemporaryFile(delete=False, suffix=".wav") as tmp:
+            speaker_path = tmp.name
+            torchaudio.save(speaker_path, waveform, sr)
+
+        # Output file
+        output_path = tempfile.NamedTemporaryFile(delete=False, suffix=".wav").name
+
+        # Generate speech
+        tts.tts_to_file(
+            text=text,
+            speaker_wav=speaker_path,
+            language=lang,
+            file_path=output_path,
+            speed=1.1   # slight speed boost
+        )
+
+        return output_path, "✅ Success"
+
+    except Exception as e:
+        return None, f"❌ Error: {str(e)}"
+
+
+# ---------------------------
+# Gradio UI
+# ---------------------------
+with gr.Blocks() as demo:
+    gr.Markdown("# 🎤 XTTS Voice Cloning (CPU Space)")
+    gr.Markdown("Upload a voice sample, enter text, choose language")
+
+    with gr.Row():
+        audio_input = gr.Audio(type="filepath", label="🎙 Sample Voice")
+        text_input = gr.Textbox(label="📝 Text", placeholder="Enter text here...")
+
+    lang_input = gr.Textbox(
+        label="🌐 Language Code",
+        value="en",
+        placeholder="en, hi, fr, de..."
+    )
+
+    generate_btn = gr.Button("🚀 Generate")
+
+    output_audio = gr.Audio(label="🔊 Output")
+    status = gr.Textbox(label="Status")
+
+    generate_btn.click(
+        fn=clone_voice,
+        inputs=[audio_input, text_input, lang_input],
+        outputs=[output_audio, status]
+    )
+
+# Required for Hugging Face Spaces
+demo.launch(server_name="0.0.0.0", server_port=7860)