Upload 3 files

src/utils/gpu-optimizer.py

@@ -0,0 +1,30 @@
+import torch
+import gc
+
+class GPUOptimizer:
+    def __init__(self):
+        self.device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+        
+    def optimize(self):
+        if torch.cuda.is_available():
+            # Clear cache
+            torch.cuda.empty_cache()
+            gc.collect()
+            
+            # Set memory fraction
+            torch.cuda.set_per_process_memory_fraction(0.9)
+            
+            # Enable TF32 for better performance
+            torch.backends.cuda.matmul.allow_tf32 = True
+            torch.backends.cudnn.allow_tf32 = True
+            
+            # Enable autocast for mixed precision
+            torch.cuda.amp.autocast(enabled=True)
+            
+    def get_memory_usage(self):
+        if torch.cuda.is_available():
+            return {
+                'allocated': torch.cuda.memory_allocated() / 1024**2,  # MB
+                'reserved': torch.cuda.memory_reserved() / 1024**2     # MB
+            }
+        return {'allocated': 0, 'reserved': 0}

src/utils/model-cache.py

@@ -0,0 +1,18 @@
+from functools import lru_cache
+import hashlib
+import json
+
+class ModelCache:
+    def __init__(self, cache_size=128):
+        self.cache_size = cache_size
+    
+    @lru_cache(maxsize=128)
+    def cache_result(self, input_key, result):
+        return result
+    
+    def get_cache_key(self, audio_data):
+        # Create hash of audio data for cache key
+        return hashlib.md5(audio_data).hexdigest()
+    
+    def clear_cache(self):
+        self.cache_result.cache_clear()

src/utils/visualizer.py

@@ -0,0 +1,74 @@
+import plotly.graph_objects as go
+from typing import Dict
+
+def create_emotion_plot(emotions: Dict[str, float]) -> str:
+    """Create emotion distribution plot"""
+    fig = go.Figure()
+    
+    # Add bar plot
+    fig.add_trace(go.Bar(
+        x=list(emotions.keys()),
+        y=list(emotions.values()),
+        marker_color='rgb(55, 83, 109)'
+    ))
+    
+    # Update layout
+    fig.update_layout(
+        title='Emotion Distribution',
+        xaxis_title='Emotion',
+        yaxis_title='Score',
+        yaxis_range=[0, 1],
+        template='plotly_white',
+        height=400
+    )
+    
+    return fig.to_html(include_plotlyjs=True)
+
+def create_pitch_plot(pitch_data: Dict) -> str:
+    """Create pitch analysis plot"""
+    fig = go.Figure()
+    
+    # Add box plot
+    fig.add_trace(go.Box(
+        y=[pitch_data['min'], pitch_data['mean'], pitch_data['max']],
+        name='Pitch Distribution',
+        boxpoints='all'
+    ))
+    
+    # Update layout
+    fig.update_layout(
+        title='Pitch Analysis',
+        yaxis_title='Frequency (Hz)',
+        template='plotly_white',
+        height=400
+    )
+    
+    return fig.to_html(include_plotlyjs=True)
+
+def create_energy_plot(energy_data: Dict) -> str:
+    """Create energy analysis plot"""
+    fig = go.Figure()
+    
+    # Add indicator
+    fig.add_trace(go.Indicator(
+        mode='gauge+number',
+        value=energy_data['mean'],
+        title={'text': 'Voice Energy Level'},
+        gauge={
+            'axis': {'range': [0, 1]},
+            'bar': {'color': 'darkblue'},
+            'steps': [
+                {'range': [0, 0.3], 'color': 'lightgray'},
+                {'range': [0.3, 0.7], 'color': 'gray'},
+                {'range': [0.7, 1], 'color': 'darkgray'}
+            ]
+        }
+    ))
+    
+    # Update layout
+    fig.update_layout(
+        height=300,
+        template='plotly_white'
+    )
+    
+    return fig.to_html(include_plotlyjs=True)