Spaces:
Paused
Paused
| import gradio as gr | |
| import torch | |
| import numpy as np | |
| import scipy.io.wavfile | |
| import tempfile | |
| import os | |
| from transformers import VitsModel, AutoTokenizer | |
| from gtts import gTTS | |
| # ============================================================ | |
| # 🎙️ Hindi Text-to-Speech Studio | |
| # Two Models: Meta MMS-TTS + Google TTS | |
| # ============================================================ | |
| # --- Global Model Loading (loaded once at startup) --- | |
| print("🔮 Loading Meta MMS-TTS Hindi model...") | |
| mms_model = VitsModel.from_pretrained("facebook/mms-tts-hin") | |
| mms_tokenizer = AutoTokenizer.from_pretrained("facebook/mms-tts-hin") | |
| print("✅ Meta MMS-TTS Hindi loaded successfully!") | |
| def generate_mms_tts(text, speed=1.0): | |
| """Generate speech using Facebook MMS-TTS Hindi (VITS model)""" | |
| if not text or not text.strip(): | |
| raise gr.Error("⚠️ कृपया कुछ हिंदी टेक्स्ट दर्ज करें! (Please enter some Hindi text!)") | |
| try: | |
| inputs = mms_tokenizer(text.strip(), return_tensors="pt") | |
| with torch.no_grad(): | |
| # Set speaking rate | |
| mms_model.speaking_rate = speed | |
| output = mms_model(**inputs) | |
| waveform = output.waveform.float().numpy().squeeze() | |
| sample_rate = mms_model.config.sampling_rate | |
| # Normalize audio | |
| if np.max(np.abs(waveform)) > 0: | |
| waveform = waveform / np.max(np.abs(waveform)) * 0.95 | |
| # Save to temp file | |
| tmp = tempfile.NamedTemporaryFile(suffix=".wav", delete=False) | |
| scipy.io.wavfile.write(tmp.name, rate=sample_rate, data=waveform) | |
| return tmp.name | |
| except Exception as e: | |
| raise gr.Error(f"❌ MMS-TTS Error: {str(e)}") | |
| def generate_google_tts(text, slow=False): | |
| """Generate speech using Google TTS Hindi""" | |
| if not text or not text.strip(): | |
| raise gr.Error("⚠️ कृपया कुछ हिंदी टेक्स्ट दर्ज करें! (Please enter some Hindi text!)") | |
| try: | |
| tts = gTTS(text=text.strip(), lang='hi', slow=slow) | |
| tmp = tempfile.NamedTemporaryFile(suffix=".mp3", delete=False) | |
| tts.save(tmp.name) | |
| return tmp.name | |
| except Exception as e: | |
| raise gr.Error(f"❌ Google TTS Error: {str(e)}") | |
| def generate_both(text, mms_speed, google_slow): | |
| """Generate speech from both models for comparison""" | |
| if not text or not text.strip(): | |
| raise gr.Error("⚠️ कृपया कुछ हिंदी टेक्स्ट दर्ज करें! (Please enter some Hindi text!)") | |
| mms_audio = generate_mms_tts(text, mms_speed) | |
| google_audio = generate_google_tts(text, google_slow) | |
| return mms_audio, google_audio | |
| # ============================================================ | |
| # 🎨 Custom CSS for Premium UI | |
| # ============================================================ | |
| custom_css = """ | |
| @import url('https://fonts.googleapis.com/css2?family=Inter:wght@300;400;500;600;700;800&family=Noto+Sans+Devanagari:wght@300;400;500;600;700&display=swap'); | |
| * { | |
| font-family: 'Inter', 'Noto Sans Devanagari', sans-serif !important; | |
| } | |
| .gradio-container { | |
| max-width: 1100px !important; | |
| margin: 0 auto !important; | |
| background: linear-gradient(135deg, #0f0c29 0%, #302b63 50%, #24243e 100%) !important; | |
| min-height: 100vh; | |
| } | |
| .main-header { | |
| text-align: center; | |
| padding: 2rem 1rem; | |
| background: linear-gradient(135deg, rgba(139, 92, 246, 0.15), rgba(59, 130, 246, 0.1)); | |
| border-radius: 20px; | |
| border: 1px solid rgba(139, 92, 246, 0.2); | |
| margin-bottom: 1.5rem; | |
| backdrop-filter: blur(10px); | |
| } | |
| .main-header h1 { | |
| font-size: 2.5rem !important; | |
| font-weight: 800 !important; | |
| background: linear-gradient(135deg, #a78bfa, #818cf8, #60a5fa) !important; | |
| -webkit-background-clip: text !important; | |
| -webkit-text-fill-color: transparent !important; | |
| background-clip: text !important; | |
| margin-bottom: 0.5rem !important; | |
| letter-spacing: -0.02em; | |
| } | |
| .main-header p { | |
| color: #a5b4fc !important; | |
| font-size: 1.1rem !important; | |
| font-weight: 400; | |
| opacity: 0.9; | |
| } | |
| .model-badge { | |
| display: inline-block; | |
| padding: 0.3rem 0.8rem; | |
| border-radius: 20px; | |
| font-size: 0.8rem; | |
| font-weight: 600; | |
| margin: 0.2rem; | |
| } | |
| .badge-meta { | |
| background: linear-gradient(135deg, #7c3aed, #6d28d9); | |
| color: white; | |
| } | |
| .badge-google { | |
| background: linear-gradient(135deg, #2563eb, #1d4ed8); | |
| color: white; | |
| } | |
| /* Tab Styling */ | |
| .tab-nav { | |
| background: rgba(30, 27, 75, 0.6) !important; | |
| border-radius: 16px !important; | |
| padding: 0.4rem !important; | |
| border: 1px solid rgba(139, 92, 246, 0.2) !important; | |
| } | |
| .tab-nav button { | |
| border-radius: 12px !important; | |
| font-weight: 600 !important; | |
| padding: 0.8rem 1.5rem !important; | |
| transition: all 0.3s ease !important; | |
| color: #a5b4fc !important; | |
| font-size: 0.95rem !important; | |
| } | |
| .tab-nav button.selected { | |
| background: linear-gradient(135deg, #7c3aed, #6d28d9) !important; | |
| color: white !important; | |
| box-shadow: 0 4px 15px rgba(124, 58, 237, 0.4) !important; | |
| } | |
| /* Input/Output Containers */ | |
| .gr-block, .gr-box, .gr-panel { | |
| background: rgba(30, 27, 75, 0.5) !important; | |
| border: 1px solid rgba(139, 92, 246, 0.15) !important; | |
| border-radius: 16px !important; | |
| backdrop-filter: blur(10px) !important; | |
| } | |
| /* Textbox */ | |
| textarea, input[type="text"] { | |
| background: rgba(15, 12, 41, 0.7) !important; | |
| border: 1px solid rgba(139, 92, 246, 0.3) !important; | |
| border-radius: 12px !important; | |
| color: #e0e7ff !important; | |
| font-size: 1.05rem !important; | |
| padding: 1rem !important; | |
| transition: all 0.3s ease !important; | |
| } | |
| textarea:focus, input[type="text"]:focus { | |
| border-color: #8b5cf6 !important; | |
| box-shadow: 0 0 20px rgba(139, 92, 246, 0.2) !important; | |
| } | |
| /* Buttons */ | |
| .gr-button-primary { | |
| background: linear-gradient(135deg, #7c3aed, #6d28d9, #5b21b6) !important; | |
| border: none !important; | |
| border-radius: 12px !important; | |
| padding: 0.8rem 2rem !important; | |
| font-weight: 700 !important; | |
| font-size: 1rem !important; | |
| color: white !important; | |
| box-shadow: 0 4px 15px rgba(124, 58, 237, 0.4) !important; | |
| transition: all 0.3s ease !important; | |
| text-transform: uppercase; | |
| letter-spacing: 0.05em; | |
| } | |
| .gr-button-primary:hover { | |
| transform: translateY(-2px) !important; | |
| box-shadow: 0 8px 25px rgba(124, 58, 237, 0.5) !important; | |
| } | |
| /* Slider */ | |
| input[type="range"] { | |
| accent-color: #8b5cf6 !important; | |
| } | |
| /* Labels */ | |
| label { | |
| color: #c4b5fd !important; | |
| font-weight: 600 !important; | |
| font-size: 0.9rem !important; | |
| } | |
| /* Audio Players */ | |
| audio { | |
| border-radius: 12px !important; | |
| width: 100% !important; | |
| } | |
| /* Checkbox */ | |
| input[type="checkbox"] { | |
| accent-color: #8b5cf6 !important; | |
| } | |
| /* Examples */ | |
| .gr-examples { | |
| background: rgba(30, 27, 75, 0.3) !important; | |
| border-radius: 16px !important; | |
| padding: 1rem !important; | |
| } | |
| /* Footer */ | |
| .footer-info { | |
| text-align: center; | |
| padding: 1.5rem; | |
| color: #6366f1; | |
| font-size: 0.85rem; | |
| opacity: 0.7; | |
| } | |
| /* Scrollbar */ | |
| ::-webkit-scrollbar { | |
| width: 6px; | |
| } | |
| ::-webkit-scrollbar-track { | |
| background: rgba(15, 12, 41, 0.3); | |
| } | |
| ::-webkit-scrollbar-thumb { | |
| background: rgba(139, 92, 246, 0.4); | |
| border-radius: 3px; | |
| } | |
| /* Animations */ | |
| @keyframes pulse-glow { | |
| 0%, 100% { box-shadow: 0 0 20px rgba(139, 92, 246, 0.2); } | |
| 50% { box-shadow: 0 0 40px rgba(139, 92, 246, 0.4); } | |
| } | |
| .generating { | |
| animation: pulse-glow 2s ease-in-out infinite; | |
| } | |
| """ | |
| # ============================================================ | |
| # 🏗️ Build the Gradio App | |
| # ============================================================ | |
| # Example Hindi texts | |
| example_texts = [ | |
| "नमस्ते! आज का मौसम बहुत सुहावना है। चलिए बाहर घूमने चलते हैं।", | |
| "भारत एक विविधताओं से भरा देश है। यहाँ अनेक भाषाएँ, संस्कृतियाँ और परंपराएँ हैं।", | |
| "कृत्रिम बुद्धिमत्ता का क्षेत्र तेजी से विकसित हो रहा है और यह हमारे जीवन को बदल रहा है।", | |
| "हिमालय की ऊँची चोटियों से लेकर कन्याकुमारी के समुद्र तट तक, भारत अपनी प्राकृतिक सुंदरता के लिए प्रसिद्ध है।", | |
| "शिक्षा सबसे शक्तिशाली हथियार है जिसका उपयोग आप दुनिया को बदलने के लिए कर सकते हैं।", | |
| "प्रौद्योगिकी ने संचार को आसान बना दिया है। अब हम दुनिया के किसी भी कोने से जुड़ सकते हैं।", | |
| ] | |
| with gr.Blocks(css=custom_css, title="🎙️ Hindi Text-to-Speech Studio", theme=gr.themes.Base()) as demo: | |
| # Header | |
| gr.HTML(""" | |
| <div class="main-header"> | |
| <h1>🎙️ हिंदी Text-to-Speech Studio</h1> | |
| <p>Premium Hindi speech synthesis powered by AI</p> | |
| <div style="margin-top: 1rem;"> | |
| <span class="model-badge badge-meta">🔮 Meta MMS-TTS</span> | |
| <span class="model-badge badge-google">🌐 Google TTS</span> | |
| </div> | |
| </div> | |
| """) | |
| with gr.Tabs() as tabs: | |
| # ═══════════════════════════════════════════ | |
| # TAB 1: Meta MMS-TTS Hindi | |
| # ═══════════════════════════════════════════ | |
| with gr.TabItem("🔮 Meta MMS-TTS Hindi", id="mms"): | |
| gr.HTML(""" | |
| <div style="padding: 1rem; text-align: center;"> | |
| <p style="color: #c4b5fd; font-size: 0.95rem;"> | |
| <strong>Facebook's Massively Multilingual Speech</strong> — Neural TTS using VITS architecture. | |
| Fast, lightweight, and produces natural-sounding Hindi speech. | |
| </p> | |
| </div> | |
| """) | |
| with gr.Row(): | |
| with gr.Column(scale=3): | |
| mms_text = gr.Textbox( | |
| label="📝 हिंदी टेक्स्ट (Hindi Text)", | |
| placeholder="यहाँ अपना हिंदी टेक्स्ट लिखें...\nExample: नमस्ते, आप कैसे हैं?", | |
| lines=5, | |
| max_lines=15, | |
| elem_id="mms-text-input" | |
| ) | |
| with gr.Column(scale=1): | |
| mms_speed = gr.Slider( | |
| minimum=0.5, | |
| maximum=2.0, | |
| value=1.0, | |
| step=0.1, | |
| label="⚡ Speaking Speed", | |
| info="0.5 = Slow, 1.0 = Normal, 2.0 = Fast" | |
| ) | |
| mms_btn = gr.Button( | |
| "🎤 Generate with MMS-TTS", | |
| variant="primary", | |
| size="lg", | |
| elem_id="mms-generate-btn" | |
| ) | |
| mms_output = gr.Audio( | |
| label="🔊 Generated Audio (Meta MMS-TTS)", | |
| type="filepath", | |
| elem_id="mms-audio-output" | |
| ) | |
| gr.Examples( | |
| examples=[[text, 1.0] for text in example_texts[:3]], | |
| inputs=[mms_text, mms_speed], | |
| label="📌 Example Hindi Texts" | |
| ) | |
| mms_btn.click( | |
| fn=generate_mms_tts, | |
| inputs=[mms_text, mms_speed], | |
| outputs=mms_output, | |
| api_name="mms_tts" | |
| ) | |
| # ═══════════════════════════════════════════ | |
| # TAB 2: Google TTS Hindi | |
| # ═══════════════════════════════════════════ | |
| with gr.TabItem("🌐 Google TTS Hindi", id="google"): | |
| gr.HTML(""" | |
| <div style="padding: 1rem; text-align: center;"> | |
| <p style="color: #93c5fd; font-size: 0.95rem;"> | |
| <strong>Google Text-to-Speech</strong> — Production-grade TTS engine. | |
| Delivers clear, natural Hindi speech with professional voice quality. | |
| </p> | |
| </div> | |
| """) | |
| with gr.Row(): | |
| with gr.Column(scale=3): | |
| google_text = gr.Textbox( | |
| label="📝 हिंदी टेक्स्ट (Hindi Text)", | |
| placeholder="यहाँ अपना हिंदी टेक्स्ट लिखें...\nExample: भारत एक महान देश है।", | |
| lines=5, | |
| max_lines=15, | |
| elem_id="google-text-input" | |
| ) | |
| with gr.Column(scale=1): | |
| google_slow = gr.Checkbox( | |
| label="🐢 Slow Speech Mode", | |
| value=False, | |
| info="Enable for slower, clearer pronunciation" | |
| ) | |
| google_btn = gr.Button( | |
| "🎤 Generate with Google TTS", | |
| variant="primary", | |
| size="lg", | |
| elem_id="google-generate-btn" | |
| ) | |
| google_output = gr.Audio( | |
| label="🔊 Generated Audio (Google TTS)", | |
| type="filepath", | |
| elem_id="google-audio-output" | |
| ) | |
| gr.Examples( | |
| examples=[[text, False] for text in example_texts[3:]], | |
| inputs=[google_text, google_slow], | |
| label="📌 Example Hindi Texts" | |
| ) | |
| google_btn.click( | |
| fn=generate_google_tts, | |
| inputs=[google_text, google_slow], | |
| outputs=google_output, | |
| api_name="google_tts" | |
| ) | |
| # ═══════════════════════════════════════════ | |
| # TAB 3: Compare Both Models | |
| # ═══════════════════════════════════════════ | |
| with gr.TabItem("⚔️ Compare Both Models", id="compare"): | |
| gr.HTML(""" | |
| <div style="padding: 1rem; text-align: center;"> | |
| <p style="color: #a5b4fc; font-size: 0.95rem;"> | |
| <strong>Side-by-Side Comparison</strong> — Enter text once and hear how both models synthesize it. | |
| Compare quality, naturalness, and pronunciation. | |
| </p> | |
| </div> | |
| """) | |
| compare_text = gr.Textbox( | |
| label="📝 हिंदी टेक्स्ट (Hindi Text)", | |
| placeholder="दोनों मॉडल की तुलना के लिए यहाँ टेक्स्ट लिखें...", | |
| lines=5, | |
| max_lines=15, | |
| elem_id="compare-text-input" | |
| ) | |
| with gr.Row(): | |
| compare_mms_speed = gr.Slider( | |
| minimum=0.5, | |
| maximum=2.0, | |
| value=1.0, | |
| step=0.1, | |
| label="⚡ MMS Speed" | |
| ) | |
| compare_google_slow = gr.Checkbox( | |
| label="🐢 Google Slow Mode", | |
| value=False | |
| ) | |
| compare_btn = gr.Button( | |
| "⚔️ Generate & Compare Both", | |
| variant="primary", | |
| size="lg", | |
| elem_id="compare-generate-btn" | |
| ) | |
| with gr.Row(): | |
| with gr.Column(): | |
| gr.HTML('<p style="text-align:center; color:#c4b5fd; font-weight:700; font-size:1.1rem;">🔮 Meta MMS-TTS</p>') | |
| compare_mms_output = gr.Audio( | |
| label="Meta MMS-TTS Output", | |
| type="filepath", | |
| elem_id="compare-mms-output" | |
| ) | |
| with gr.Column(): | |
| gr.HTML('<p style="text-align:center; color:#93c5fd; font-weight:700; font-size:1.1rem;">🌐 Google TTS</p>') | |
| compare_google_output = gr.Audio( | |
| label="Google TTS Output", | |
| type="filepath", | |
| elem_id="compare-google-output" | |
| ) | |
| gr.Examples( | |
| examples=[[text, 1.0, False] for text in example_texts[:4]], | |
| inputs=[compare_text, compare_mms_speed, compare_google_slow], | |
| label="📌 Try These Examples" | |
| ) | |
| compare_btn.click( | |
| fn=generate_both, | |
| inputs=[compare_text, compare_mms_speed, compare_google_slow], | |
| outputs=[compare_mms_output, compare_google_output], | |
| api_name="compare_both" | |
| ) | |
| # Footer | |
| gr.HTML(""" | |
| <div class="footer-info"> | |
| <p>🎙️ Hindi Text-to-Speech Studio | Powered by Meta MMS & Google TTS</p> | |
| <p style="margin-top: 0.3rem; font-size: 0.75rem;"> | |
| Models: facebook/mms-tts-hin (VITS) • gTTS (Google) | |
| </p> | |
| </div> | |
| """) | |
| # Launch | |
| if __name__ == "__main__": | |
| demo.launch() | |