Expand description
Cross-modal embeddings for multi-modal vector search
This module provides CLIP-style cross-modal embeddings that can handle:
- Text-image alignment
- Multi-modal fusion
- Cross-modal attention mechanisms
- Joint embedding spaces
Structs§
- Attention
Mechanism - Attention mechanism for cross-modal alignment
- Audio
Data - Audio data representation
- Cross
Modal Config - Configuration for cross-modal embeddings
- Cross
Modal Encoder - Cross-modal embedding encoder that handles multiple modalities
- Fusion
Layer - Fusion layer for combining multiple modalities
- Graph
Data - Graph data representation for knowledge graphs
- Graph
Edge - Graph
Node - Image
Data - Image data representation
- Mock
Audio Encoder - Mock
Graph Encoder - Mock
Image Encoder - Similar mock implementations for other modalities
- Mock
Text Encoder - Simple implementations of encoder traits for testing
- Mock
Video Encoder - Multi
Modal Content - Multi-modal content that can contain multiple types of data
- Spatial
Info - Spatial information for location-aware embeddings
- Temporal
Info - Temporal information for time-series data
- Video
Data - Video data representation
Enums§
- Fusion
Strategy - Fusion strategies for combining multiple modalities
- Image
Format - Modality
- Modality types supported by the cross-modal system
- Modality
Data - Data for a specific modality
Traits§
- Audio
Encoder - Audio encoder trait for cross-modal systems
- Graph
Encoder - Graph encoder trait for knowledge graph embeddings
- Image
Encoder - Image encoder trait for cross-modal systems
- Text
Encoder - Text encoder trait for cross-modal systems
- Video
Encoder - Video encoder trait for cross-modal systems