Switch to Whisper base

for faster voice processing

- Change from small (20s) to base
(~5-8s) model
- Better CPU performance while
maintaining good accuracy
- Reduces total voice query time from
20s to 8-10s

README.md

@@ -92,7 +92,7 @@ Priority given to division accuracy:
 ### Models Used
 - **Embeddings**: `sentence-transformers/all-MiniLM-L6-v2` - Fast, lightweight semantic search
 - **Name Extraction**: `urchade/gliner_small-v2.1` - Zero-shot NER for person names
-- **Speech-to-Text**: `openai/whisper-small` - Better accuracy for multi-language transcription
+- **Speech-to-Text**: `openai/whisper-base` - Optimized for CPU with good accuracy
 
 ### Confidence Scoring
 

app.py

@@ -41,9 +41,9 @@ def initialize_services():
     name_extractor = NameExtractor(model_name="urchade/gliner_small-v2.1")
     logger.info("✓ Name extractor ready!")
 
-    # Initialize voice processor (using small model for better accuracy)
+    # Initialize voice processor (using base model for speed/accuracy balance)
     logger.info("Loading Whisper model...")
-    voice_processor = VoiceProcessor(model_size="small")  # Using small for better accuracy
+    voice_processor = VoiceProcessor(model_size="base")  # Using base for best speed/accuracy on CPU
     logger.info("✓ Voice processor ready!")
 
     # Initialize contact search