Upload 12 files

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+qari[[:space:]]part-1.mp4 filter=lfs diff=lfs merge=lfs -text

app.py

@@ -0,0 +1,783 @@
+import streamlit as st
+from google.oauth2 import service_account
+from google.cloud import speech
+import io
+import torch
+import numpy as np
+from transformers import Wav2Vec2Processor
+from transformers.models.wav2vec2 import Wav2Vec2Model
+import librosa
+from groq import Groq
+import sounddevice as sd
+import scipy.io.wavfile as wav
+import os
+from datetime import datetime
+from pydub import AudioSegment
+from pathlib import Path
+from openai import OpenAI
+import json
+import plotly.graph_objects as go
+import os
+
+OpenAI_api_key =os.environ.get('OpenAI_api_key')
+
+Groq_api_key = os.environ.get('Groq_api_key')
+
+google_creds = os.environ.get('google_creds')
+
+# Enhanced UI Styles
+CUSTOM_CSS = """
+<style>
+@import url('https://fonts.googleapis.com/css2?family=Inter:wght@400;500;600;700&display=swap');
+
+/* Base styles */
+:root {
+    --primary-color: #2563eb;
+    --secondary-color: #1d4ed8;
+    --success-color: #059669;
+    --warning-color: #d97706;
+    --danger-color: #dc2626;
+    --text-primary: #111827;
+    --text-secondary: #4b5563;
+    --bg-primary: #ffffff;
+    --bg-secondary: #f3f4f6;
+}
+
+.stApp {
+    font-family: 'Inter', sans-serif;
+    color: var(--text-primary);
+    background: var(--bg-secondary);
+}
+
+/* Header styles */
+.app-header {
+    background: linear-gradient(135deg, var(--primary-color), var(--secondary-color));
+    padding: 2rem 1rem;
+    text-align: center;
+    border-radius: 0 0 1.5rem 1.5rem;
+    margin-bottom: 2rem;
+    box-shadow: 0 4px 6px rgba(0, 0, 0, 0.1);
+}
+
+.app-title {
+    color: white;
+    font-size: 2.5rem;
+    font-weight: 700;
+    margin-bottom: 0.5rem;
+    text-shadow: 0 2px 4px rgba(0, 0, 0, 0.1);
+}
+
+.app-subtitle {
+    color: rgba(255, 255, 255, 0.9);
+    font-size: 1.2rem;
+    font-weight: 500;
+    direction: rtl;
+}
+
+/* Card styles */
+.card {
+    background: var(--bg-primary);
+    border-radius: 1rem;
+    padding: 1.5rem;
+    margin-bottom: 1.5rem;
+    box-shadow: 0 4px 6px rgba(0, 0, 0, 0.1);
+    transition: transform 0.2s ease;
+}
+
+.card:hover {
+    transform: translateY(-2px);
+}
+
+.card-header {
+    display: flex;
+    align-items: center;
+    gap: 0.75rem;
+    margin-bottom: 1rem;
+    padding-bottom: 0.75rem;
+    border-bottom: 1px solid var(--bg-secondary);
+}
+
+.card-title {
+    font-size: 1.25rem;
+    font-weight: 600;
+    color: var(--text-primary);
+    margin: 0;
+}
+
+/* Button styles */
+.button-container {
+    display: flex;
+    gap: 1rem;
+    margin-bottom: 1rem;
+}
+
+.button-primary {
+    background-color: var(--primary-color);
+    color: white;
+    padding: 0.75rem 1.5rem;
+    border-radius: 0.5rem;
+    border: none;
+    font-weight: 500;
+    cursor: pointer;
+    transition: background-color 0.2s ease;
+    text-align: center;
+    display: inline-flex;
+    align-items: center;
+    justify-content: center;
+    gap: 0.5rem;
+}
+
+.button-primary:hover {
+    background-color: var(--secondary-color);
+}
+
+.button-danger {
+    background-color: var(--danger-color);
+    color: white;
+    padding: 0.75rem 1.5rem;
+    border-radius: 0.5rem;
+    border: none;
+    font-weight: 500;
+    cursor: pointer;
+    transition: background-color 0.2s ease;
+}
+
+/* Progress indicator */
+.score-container {
+    text-align: center;
+    padding: 1.5rem;
+    background: var(--bg-secondary);
+    border-radius: 1rem;
+    margin-bottom: 1.5rem;
+}
+
+.score-value {
+    font-size: 3rem;
+    font-weight: 700;
+    color: var(--primary-color);
+}
+
+.score-label {
+    color: var(--text-secondary);
+    font-size: 1.1rem;
+    margin-top: 0.5rem;
+}
+
+/* Feedback section */
+.feedback-section {
+    background: var(--bg-secondary);
+    border-radius: 1rem;
+    padding: 1.5rem;
+    margin-top: 1.5rem;
+}
+
+.feedback-item {
+    background: white;
+    border-radius: 0.5rem;
+    padding: 1rem;
+    margin-bottom: 1rem;
+    box-shadow: 0 2px 4px rgba(0, 0, 0, 0.05);
+}
+
+/* Status messages */
+.success-msg {
+    background-color: var(--success-color);
+    color: white;
+    padding: 1rem;
+    border-radius: 0.5rem;
+    text-align: center;
+    margin-bottom: 1rem;
+    animation: slideIn 0.3s ease;
+}
+
+.error-msg {
+    background-color: var(--danger-color);
+    color: white;
+    padding: 1rem;
+    border-radius: 0.5rem;
+    text-align: center;
+    margin-bottom: 1rem;
+    animation: slideIn 0.3s ease;
+}
+
+/* Animations */
+@keyframes slideIn {
+    from { transform: translateY(-10px); opacity: 0; }
+    to { transform: translateY(0); opacity: 1; }
+}
+
+/* Responsive adjustments */
+@media (max-width: 768px) {
+    .app-title {
+        font-size: 2rem;
+    }
+    
+    .card {
+        padding: 1rem;
+    }
+    
+    .button-container {
+        flex-direction: column;
+    }
+    
+    .score-value {
+        font-size: 2.5rem;
+    }
+}
+</style>
+"""
+
+
+class AzanTrainerApp:
+    def __init__(self):
+        self.setup_api_clients()
+        self.setup_configs()
+        self.setup_directories()
+        self.initialize_models()
+
+    def setup_api_clients(self):
+        """Initialize API clients"""
+        self.openai_client = OpenAI(api_key=OpenAI_api_key)
+        self.groq_client = Groq(api_key=Groq_api_key)
+        self.speech_client = self.init_google_speech()
+
+    def init_google_speech(self):
+        """Initialize Google Speech client"""
+        credentials = service_account.Credentials.from_service_account_file(google_creds)
+        return speech.SpeechClient(credentials=credentials)
+
+    def setup_configs(self):
+        """Set up configuration variables"""
+        self.SAMPLE_RATE = 48000
+        self.DURATION = 6
+        self.AUDIO_GAIN = 1.50
+        self.IDEAL_TEXT = "اللّٰهُ أَكْبَرُ، اللّٰهُ أَكْبَرُ"
+        self.IDEAL_TEXT_MEANING = "Allah is the Greatest, Allah is the Greatest"
+
+    def setup_directories(self):
+        """Create necessary directories"""
+        for dir_name in ['recordings', 'feedback_audio']:
+            os.makedirs(dir_name, exist_ok=True)
+
+    def initialize_models(self):
+        """Initialize ML models"""
+        self.processor = Wav2Vec2Processor.from_pretrained("models/wav2vec2-base")
+        self.model = Wav2Vec2Model.from_pretrained("models/wav2vec2-base")
+        self.ideal_embedding = torch.tensor(np.load("ideal_embedding_part_1.npy"))
+
+    def create_waveform_visualization(self, audio_path, reference_path):
+        """Create waveform visualization using Plotly"""
+        fig = go.Figure()
+
+        # Process user audio
+        y_user, sr_user = librosa.load(audio_path)
+        times_user = np.arange(len(y_user)) / sr_user
+        fig.add_trace(go.Scatter(
+            x=times_user,
+            y=y_user,
+            name='Your Recording',
+            line=dict(color='#1E88E5')
+        ))
+
+        # Process reference audio
+        y_ref, sr_ref = librosa.load(reference_path)
+        times_ref = np.arange(len(y_ref)) / sr_ref
+        fig.add_trace(go.Scatter(
+            x=times_ref,
+            y=y_ref,
+            name='Expert Recording',
+            line=dict(color='#4CAF50')
+        ))
+
+        fig.update_layout(
+            title='Waveform Comparison',
+            xaxis_title='Time (s)',
+            yaxis_title='Amplitude',
+            template='plotly_white',
+            height=400
+        )
+
+        return fig
+
+    def record_audio(self):
+        """Record audio from user"""
+        try:
+            audio_data = sd.rec(
+                int(self.DURATION * self.SAMPLE_RATE),
+                samplerate=self.SAMPLE_RATE,
+                channels=1,
+                dtype=np.float32
+            )
+            sd.wait()
+            return self.enhance_audio(audio_data)
+        except Exception as e:
+            st.error(f"Recording error: {str(e)}")
+            return None
+
+    def enhance_audio(self, audio_data):
+        """Enhance audio quality"""
+        audio_data = audio_data / (np.max(np.abs(audio_data)) + 1e-10)
+        audio_data = audio_data * self.AUDIO_GAIN
+        noise_threshold = 0.01
+        audio_data[np.abs(audio_data) < noise_threshold] = 0
+        return audio_data
+
+    def save_audio(self, audio_data):
+        """Save audio to file"""
+        timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
+        filename = f"recordings/audio_{timestamp}.wav"
+        audio_data = np.clip(audio_data * 32767, -32768, 32767).astype(np.int16)
+        wav.write(filename, self.SAMPLE_RATE, audio_data)
+        return filename
+
+    def analyze_recording(self, audio_path):
+        """Analyze the recording"""
+        try:
+            # Convert to MP3 for Google Speech API
+            mp3_path = audio_path.replace('.wav', '.mp3')
+            AudioSegment.from_wav(audio_path).export(mp3_path, format="mp3")
+
+            # Transcribe audio
+            with open(mp3_path, 'rb') as f:
+                content = f.read()
+
+            audio = speech.RecognitionAudio(content=content)
+            config = speech.RecognitionConfig(
+                encoding=speech.RecognitionConfig.AudioEncoding.MP3,
+                sample_rate_hertz=self.SAMPLE_RATE,
+                language_code="ar"
+            )
+
+            response = self.speech_client.recognize(config=config, audio=audio)
+            transcription = " ".join(result.alternatives[0].transcript 
+                                   for result in response.results)
+
+            # Calculate similarity
+            user_embedding = self.get_audio_embedding(audio_path)
+            similarity_score = self.calculate_similarity(user_embedding, self.ideal_embedding)
+
+            # Generate feedback
+            feedback = self.generate_feedback(transcription, similarity_score)
+
+            # Clean up
+            os.remove(mp3_path)
+
+            return transcription, similarity_score, feedback
+
+        except Exception as e:
+            st.error(f"Analysis error: {str(e)}")
+            return None, None, None
+
+    def get_audio_embedding(self, audio_path):
+        """Generate audio embedding"""
+        audio_input, _ = librosa.load(audio_path, sr=16000)
+        inputs = self.processor(audio_input, sampling_rate=16000, 
+                              return_tensors="pt", padding=True)
+        with torch.no_grad():
+            embedding = self.model(inputs.input_values).last_hidden_state.mean(dim=1).squeeze()
+        return embedding
+
+    def calculate_similarity(self, embedding1, embedding2):
+        """Calculate similarity score"""
+        similarity = torch.nn.functional.cosine_similarity(embedding1, embedding2, dim=0)
+        return similarity.item() * 100
+
+    def generate_feedback(self, transcription, similarity_score):
+        """Generate feedback in natural Roman Urdu using LLM"""
+        prompt = f"""
+        Is Azan ki tilawat ka jaiza len aur natural Roman Urdu main feedback den:
+
+        Tilawat: {transcription}
+        Mutabiqat Score: {similarity_score:.2f}%
+
+        Feedback ko in 3 hisson main takseem karen:
+
+        1. Talaffuz (Pronunciation):
+        - Har lafz ka talaffuz kaisa hai
+        - Huroof ki tartib theek hai ya nahi
+        - Allah ke lafz ka talaffuz kaisa hai
+        - Mukammal Azan ki tarteeb kaisi hai
+
+        2. Waqt aur Lehja (Timing):
+        - Har hissay ka sahi dohrao
+        - Waqfay ki durustagi
+        - Aawaz ka utaar chadhao
+
+        3. Behtar Karne Ke Liye Mashwaray:
+        - Kahan ghaltiyan hain
+        - Kya behtar karna hai
+        - Kis cheez par zyada mehnat ki zaroorat hai
+
+        Note: Feedback zabaan-e-urdu main likhen, lekin English huroof istimal karen. 
+        Lehja mohtaram aur madadgaar hona chahiye.
+        """
+
+        response = self.groq_client.chat.completions.create(
+            model="llama3-70b-8192",
+            messages=[{"role": "user", "content": prompt}],
+            temperature=0.7,
+            max_tokens=1000
+        )
+
+        return response.choices[0].message.content
+
+    def generate_audio_feedback(self, feedback_text):
+        """Generate audio feedback"""
+        try:
+            timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
+            audio_path = f"feedback_audio/feedback_{timestamp}.mp3"
+
+            response = self.openai_client.audio.speech.create(
+                model="tts-1",
+                voice="alloy",
+                input=feedback_text
+            )
+
+            response.stream_to_file(audio_path)
+            return audio_path
+
+        except Exception as e:
+            st.error(f"Error generating audio feedback: {str(e)}")
+            return None
+
+    def run(self):
+        """Run the enhanced Streamlit application with Persian/Masjid-inspired UI"""
+        st.set_page_config(
+            page_title="Azan Pronunciation Trainer",
+            layout="wide",
+            initial_sidebar_state="collapsed"
+        )
+        
+        # Custom CSS with Persian/Masjid-inspired theme (Keep your existing CSS here)
+        st.markdown("""
+            <style>
+            /* Global Styles */
+            @import url('https://fonts.googleapis.com/css2?family=Amiri:wght@400;700&display=swap');
+            
+            :root {
+                --primary-color: #1F4C6B;
+                --secondary-color: #C3934B;
+                --accent-color: #E6B17E;
+                --background-color: #F7F3E9;
+                --text-color: #2C3E50;
+                --card-shadow: 0 4px 6px rgba(0, 0, 0, 0.1);
+            }
+
+            .stApp {
+                background-color: var(--background-color);
+                font-family: 'Amiri', serif;
+            }
+
+            /* Header Styles */
+            .app-header {
+                background: linear-gradient(135deg, var(--primary-color), #2C3E50);
+                color: white;
+                padding: 2rem;
+                border-radius: 15px;
+                text-align: center;
+                margin-bottom: 2rem;
+                box-shadow: var(--card-shadow);
+            }
+
+            .app-title {
+                font-size: 2.5rem;
+                margin-bottom: 0.5rem;
+                font-weight: 700;
+                background: linear-gradient(45deg, var(--accent-color), #FFD700);
+                -webkit-background-clip: text;
+                -webkit-text-fill-color: transparent;
+            }
+
+            .app-subtitle {
+                font-size: 1.2rem;
+                opacity: 0.9;
+                margin: 0.5rem 0;
+            }
+
+            .arabic-text {
+                font-family: 'Amiri', serif;
+                font-size: 2rem;
+                direction: rtl;
+                margin: 1rem 0;
+                color: var(--secondary-color);
+            }
+
+            /* Card Styles */
+            .card {
+                background: white;
+                border-radius: 15px;
+                padding: 1.5rem;
+                margin-bottom: 1.5rem;
+                box-shadow: var(--card-shadow);
+                border: 1px solid rgba(195, 147, 75, 0.2);
+                transition: transform 0.2s ease;
+            }
+
+            .card:hover {
+                transform: translateY(-2px);
+            }
+
+            .card-header {
+                display: flex;
+                align-items: center;
+                margin-bottom: 1rem;
+                border-bottom: 2px solid var(--accent-color);
+                padding-bottom: 0.5rem;
+            }
+
+            .card-title {
+                font-size: 1.3rem;
+                margin: 0 0 0 0.5rem;
+                color: var(--primary-color);
+            }
+
+            /* Button Styles */
+            .stButton button {
+                background: linear-gradient(45deg, var(--primary-color), var(--secondary-color));
+                color: white;
+                border: none;
+                padding: 0.75rem 1.5rem;
+                border-radius: 25px;
+                font-weight: bold;
+                transition: all 0.3s ease;
+                width: 100%;
+                margin: 0.5rem 0;
+            }
+
+            .stButton button:hover {
+                transform: translateY(-2px);
+                box-shadow: 0 4px 12px rgba(31, 76, 107, 0.2);
+            }
+
+            /* Score Display */
+            .score-container {
+                background: linear-gradient(135deg, var(--primary-color), var(--secondary-color));
+                color: white;
+                padding: 2rem;
+                border-radius: 15px;
+                text-align: center;
+                margin: 1.5rem 0;
+            }
+
+            .score-value {
+                font-size: 3rem;
+                font-weight: bold;
+                margin-bottom: 0.5rem;
+            }
+
+            .score-label {
+                font-size: 1.2rem;
+                opacity: 0.9;
+            }
+
+            /* Feedback Styles */
+            .feedback-item {
+                background-color: rgba(195, 147, 75, 0.1);
+                padding: 1rem;
+                border-radius: 10px;
+                margin: 1rem 0;
+                border-left: 4px solid var(--secondary-color);
+            }
+
+            /* Help Section Styling */
+            .help-container {
+                background: white;
+                padding: 1.5rem;
+                border-radius: 15px;
+                margin-top: 1rem;
+            }
+
+            .help-item {
+                display: flex;
+                align-items: center;
+                margin-bottom: 1rem;
+                padding: 0.5rem;
+                border-radius: 8px;
+                background-color: rgba(31, 76, 107, 0.05);
+            }
+
+            .help-number {
+                background-color: var(--primary-color);
+                color: white;
+                width: 24px;
+                height: 24px;
+                border-radius: 50%;
+                display: flex;
+                align-items: center;
+                justify-content: center;
+                margin-right: 1rem;
+                font-size: 0.9rem;
+            }
+            </style>
+        """, unsafe_allow_html=True)
+
+        # Enhanced Header with Arabic Styling
+        st.markdown(f"""
+            <div class="app-header">
+                <h1 class="app-title">Azan Pronunciation Trainer</h1>
+                <p class="app-subtitle">Perfect Your Recitation</p>
+                <div class="arabic-text">{self.IDEAL_TEXT}</div>
+                <p class="app-subtitle">{self.IDEAL_TEXT_MEANING}</p>
+            </div>
+        """, unsafe_allow_html=True)
+
+        # Expert demonstration card
+        st.markdown("""
+            <div class="card">
+                <div class="card-header">
+                    <span style="font-size: 2rem;">📹</span>
+                    <h2 class="card-title">Expert Demonstration</h2>
+                </div>
+            """, unsafe_allow_html=True)
+        st.video("qari part-1.mp4")
+        st.markdown("</div>", unsafe_allow_html=True)
+
+        # Expert audio card
+        st.markdown("""
+            <div class="card">
+                <div class="card-header">
+                    <span style="font-size: 2rem;">🎵</span>
+                    <h2 class="card-title">Reference Audio</h2>
+                </div>
+            """, unsafe_allow_html=True)
+        st.audio("qari_part_1.mp3")
+        st.markdown("</div>", unsafe_allow_html=True)
+
+        # Recording controls card
+        st.markdown("""
+            <div class="card">
+                <div class="card-header">
+                    <span style="font-size: 2rem;">🎙️</span>
+                    <h2 class="card-title">Recording Controls</h2>
+                </div>
+            """, unsafe_allow_html=True)
+        
+        col1, col2 = st.columns(2)
+        
+        with col1:
+            if st.button("Start Recording", help="Click to start recording (6 seconds)", key="start_rec"):
+                with st.spinner("Recording in progress..."):
+                    audio_data = self.record_audio()
+                    if audio_data is not None:
+                        audio_path = self.save_audio(audio_data)
+                        st.session_state['audio_file'] = audio_path
+                        st.markdown("""
+                            <div class="feedback-item" style="background-color: rgba(46, 204, 113, 0.1); border-left-color: #2ecc71;">
+                                Recording completed successfully! ✅
+                            </div>
+                        """, unsafe_allow_html=True)
+
+        with col2:
+            if st.button("Clear Recording", key="clear_rec"):
+                if 'audio_file' in st.session_state:
+                    if os.path.exists(st.session_state['audio_file']):
+                        os.remove(st.session_state['audio_file'])
+                    st.session_state['audio_file'] = None
+                    st.markdown("""
+                        <div class="feedback-item" style="background-color: rgba(231, 76, 60, 0.1); border-left-color: #e74c3c;">
+                            Recording cleared! 🗑️
+                        </div>
+                    """, unsafe_allow_html=True)
+        
+        st.markdown("</div>", unsafe_allow_html=True)
+
+        # Analysis section
+        if 'audio_file' in st.session_state and st.session_state['audio_file']:
+            st.markdown("""
+                <div class="card">
+                    <div class="card-header">
+                        <span style="font-size: 2rem;">🎵</span>
+                        <h2 class="card-title">Your Recording</h2>
+                    </div>
+                """, unsafe_allow_html=True)
+            
+            st.audio(st.session_state['audio_file'])
+
+            if st.button("Analyze Recording", key="analyze"):
+                with st.spinner("Analyzing your recitation..."):
+                    transcription, similarity, feedback = self.analyze_recording(
+                        st.session_state['audio_file']
+                    )
+
+                    if all([transcription, similarity, feedback]):
+                        # Enhanced similarity score display
+                        st.markdown(f"""
+                            <div class="score-container">
+                                <div class="score-value">{similarity:.1f}%</div>
+                                <div class="score-label">Similarity Score</div>
+                            </div>
+                        """, unsafe_allow_html=True)
+
+                        # Waveform visualization
+                        fig = self.create_waveform_visualization(
+                            st.session_state['audio_file'],
+                            "qari_part_1.mp3"
+                        )
+                        st.plotly_chart(fig, use_container_width=True)
+
+                        # Feedback display
+                        st.markdown(f"""
+                            <div class="card">
+                                <div class="card-header">
+                                    <span style="font-size: 2rem;">📝</span>
+                                    <h2 class="card-title">Detailed Feedback</h2>
+                                </div>
+                                <div class="feedback-item">
+                                    {feedback}
+                                </div>
+                            </div>
+                        """, unsafe_allow_html=True)
+
+                        # Audio feedback
+                        audio_feedback_path = self.generate_audio_feedback(feedback)
+                        if audio_feedback_path:
+                            st.markdown("""
+                                <div class="card">
+                                    <div class="card-header">
+                                        <span style="font-size: 2rem;">🔊</span>
+                                        <h2 class="card-title">Audio Feedback</h2>
+                                    </div>
+                                """, unsafe_allow_html=True)
+                            st.audio(audio_feedback_path)
+                            st.markdown("</div>", unsafe_allow_html=True)
+            
+            st.markdown("</div>", unsafe_allow_html=True)
+
+        # Enhanced help section with numbered steps
+        with st.expander("❓ How to Use"):
+            st.markdown("""
+                <div class="help-container">
+                    <div class="help-item">
+                        <div class="help-number">1</div>
+                        <div>Watch the expert demonstration video carefully</div>
+                    </div>
+                    <div class="help-item">
+                        <div class="help-number">2</div>
+                        <div>Listen to the reference audio to understand proper pronunciation</div>
+                    </div>
+                    <div class="help-item">
+                        <div class="help-number">3</div>
+                        <div>Click 'Start Recording' and recite the phrase (6 seconds)</div>
+                    </div>
+                    <div class="help-item">
+                        <div class="help-number">4</div>
+                        <div>Wait for the recording to complete</div>
+                    </div>
+                    <div class="help-item">
+                        <div class="help-number">5</div>
+                        <div>Click 'Analyze Recording' to get detailed feedback</div>
+                    </div>
+                    <div class="help-item">
+                        <div class="help-number">6</div>
+                        <div>Review your score and feedback to improve</div>
+                    </div>
+                    <div class="help-item">
+                        <div class="help-number">7</div>
+                        <div>Practice until you achieve 90% or higher similarity</div>
+                    </div>
+                </div>
+            """, unsafe_allow_html=True)
+
+if __name__ == "__main__":
+    app = AzanTrainerApp()
+    app.run()

ideal_embedding_part_1.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5d9ccaf726d2a0038435d0254359f5144a30af97cf81de5f309e7dc3f519fc67
+size 3200

models/wav2vec2-base/README.md

@@ -0,0 +1,28 @@
+---
+language: en
+datasets:
+- librispeech_asr
+tags:
+- speech
+license: apache-2.0
+---
+
+# Wav2Vec2-Base 
+
+[Facebook's Wav2Vec2](https://ai.facebook.com/blog/wav2vec-20-learning-the-structure-of-speech-from-raw-audio/)
+
+The base model pretrained on 16kHz sampled speech audio. When using the model make sure that your speech input is also sampled at 16Khz. 
+
+**Note**: This model does not have a tokenizer as it was pretrained on audio alone. In order to use this model **speech recognition**, a tokenizer should be created and the model should be fine-tuned on labeled text data. Check out [this blog](https://huggingface.co/blog/fine-tune-wav2vec2-english) for more in-detail explanation of how to fine-tune the model.
+
+[Paper](https://arxiv.org/abs/2006.11477)
+
+Authors: Alexei Baevski, Henry Zhou, Abdelrahman Mohamed, Michael Auli
+
+**Abstract**
+We show for the first time that learning powerful representations from speech audio alone followed by fine-tuning on transcribed speech can outperform the best semi-supervised methods while being conceptually simpler. wav2vec 2.0 masks the speech input in the latent space and solves a contrastive task defined over a quantization of the latent representations which are jointly learned. Experiments using all labeled data of Librispeech achieve 1.8/3.3 WER on the clean/other test sets. When lowering the amount of labeled data to one hour, wav2vec 2.0 outperforms the previous state of the art on the 100 hour subset while using 100 times less labeled data. Using just ten minutes of labeled data and pre-training on 53k hours of unlabeled data still achieves 4.8/8.2 WER. This demonstrates the feasibility of speech recognition with limited amounts of labeled data.
+The original model can be found under https://github.com/pytorch/fairseq/tree/master/examples/wav2vec#wav2vec-20.
+
+# Usage
+
+See [this notebook](https://colab.research.google.com/drive/1FjTsqbYKphl9kL-eILgUc-bl4zVThL8F?usp=sharing) for more information on how to fine-tune the model.

models/wav2vec2-base/config.json

@@ -0,0 +1,85 @@
+{
+  "activation_dropout": 0.0,
+  "apply_spec_augment": true,
+  "architectures": [
+    "Wav2Vec2ForPreTraining"
+  ],
+  "attention_dropout": 0.1,
+  "bos_token_id": 1,
+  "codevector_dim": 256,
+  "contrastive_logits_temperature": 0.1,
+  "conv_bias": false,
+  "conv_dim": [
+    512,
+    512,
+    512,
+    512,
+    512,
+    512,
+    512
+  ],
+  "conv_kernel": [
+    10,
+    3,
+    3,
+    3,
+    3,
+    2,
+    2
+  ],
+  "conv_stride": [
+    5,
+    2,
+    2,
+    2,
+    2,
+    2,
+    2
+  ],
+  "ctc_loss_reduction": "sum",
+  "ctc_zero_infinity": false,
+  "diversity_loss_weight": 0.1,
+  "do_stable_layer_norm": false,
+  "eos_token_id": 2,
+  "feat_extract_activation": "gelu",
+  "feat_extract_norm": "group",
+  "feat_proj_dropout": 0.1,
+  "feat_quantizer_dropout": 0.0,
+  "final_dropout": 0.0,
+  "freeze_feat_extract_train": true,
+  "gradient_checkpointing": true,
+  "hidden_act": "gelu",
+  "hidden_dropout": 0.1,
+  "hidden_size": 768,
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "layer_norm_eps": 1e-05,
+  "layerdrop": 0.0,
+  "mask_channel_length": 10,
+  "mask_channel_min_space": 1,
+  "mask_channel_other": 0.0,
+  "mask_channel_prob": 0.0,
+  "mask_channel_selection": "static",
+  "mask_feature_length": 10,
+  "mask_feature_prob": 0.0,
+  "mask_time_length": 10,
+  "mask_time_min_space": 1,
+  "mask_time_other": 0.0,
+  "mask_time_prob": 0.05,
+  "mask_time_selection": "static",
+  "model_type": "wav2vec2",
+  "no_mask_channel_overlap": false,
+  "no_mask_time_overlap": false,
+  "num_attention_heads": 12,
+  "num_codevector_groups": 2,
+  "num_codevectors_per_group": 320,
+  "num_conv_pos_embedding_groups": 16,
+  "num_conv_pos_embeddings": 128,
+  "num_feat_extract_layers": 7,
+  "num_hidden_layers": 12,
+  "num_negatives": 100,
+  "pad_token_id": 0,
+  "proj_codevector_dim": 256,
+  "transformers_version": "4.7.0.dev0",
+  "vocab_size": 32
+}

models/wav2vec2-base/gitattributes

@@ -0,0 +1,17 @@
+*.bin.* filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tar.gz filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text

models/wav2vec2-base/preprocessor_config.json

@@ -0,0 +1,8 @@
+{
+  "do_normalize": true,
+  "feature_size": 1,
+  "padding_side": "right",
+  "padding_value": 0.0,
+  "return_attention_mask": false,
+  "sampling_rate": 16000
+}

models/wav2vec2-base/pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3249fe98bfc62fcbc26067f724716a6ec49d12c4728a2af1df659013905dff21
+size 380267417

models/wav2vec2-base/special_tokens_map.json

@@ -0,0 +1 @@
+{"bos_token": "<s>", "eos_token": "</s>", "unk_token": "<unk>", "pad_token": "<pad>"}

models/wav2vec2-base/tokenizer_config.json

@@ -0,0 +1 @@
+{"unk_token": "<unk>", "bos_token": "<s>", "eos_token": "</s>", "pad_token": "<pad>", "do_lower_case": false, "return_attention_mask": false, "do_normalize": true}

models/wav2vec2-base/vocab.json

@@ -0,0 +1 @@
+{"<pad>": 0, "<s>": 1, "</s>": 2, "<unk>": 3, "|": 4, "E": 5, "T": 6, "A": 7, "O": 8, "N": 9, "I": 10, "H": 11, "S": 12, "R": 13, "D": 14, "L": 15, "U": 16, "M": 17, "W": 18, "C": 19, "F": 20, "G": 21, "Y": 22, "P": 23, "B": 24, "V": 25, "K": 26, "'": 27, "X": 28, "J": 29, "Q": 30, "Z": 31}

qari part-1.mp4

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9503395f4bac517d0be5011bdd4558b30e2ae6257f6dbdc0b69da716622e2d86
+size 150502392