Expand description
Native FAISS Integration with Real Bindings
This module provides actual integration with Facebook’s FAISS library through native bindings, enabling high-performance vector search with full FAISS capabilities.
Features:
- Real FAISS index import/export
- Native FAISS performance optimization
- GPU acceleration integration
- Memory-efficient batch processing
- Performance benchmarking against FAISS
Structs§
- Allocation
Stats - Memory allocation statistics
- Benchmark
Dataset - Benchmark dataset
- Benchmark
Result - Individual benchmark result
- Comparison
Data - Performance comparison data
- Comparison
Result - Comparison result
- Dataset
Characteristics - Dataset characteristics for benchmarking
- Faiss
Performance Comparison - Performance comparison framework
- GpuContext
- GPU context for FAISS GPU operations
- GpuMetrics
- GPU performance metrics
- GpuResource
- GPU resource handle
- Memory
Block - Memory block in the pool
- Memory
Metrics - Memory efficiency metrics
- Memory
Pool - Memory pool for efficient memory management
- Native
Faiss Config - Native FAISS integration configuration
- Native
Faiss Index - Native FAISS index wrapper
- Native
Faiss Statistics - Native FAISS statistics with detailed metrics
- Native
Metrics - Native FAISS specific metrics
- Native
Performance Tuning - Native performance tuning configuration
- Performance
Metrics - Performance metrics for comparison
- Performance
Ratios - Performance ratios
- Statistical
Significance - Statistical significance test results