| Property | Value |
|---|---|
| Parameter Count | 33.4M |
| License | MIT |
| Downloads | 123,821 |
| Language | English |
| Framework Support | PyTorch, ONNX |
What is bge-small-en-v1.5?
bge-small-en-v1.5 is a lightweight embedding model designed for efficient sentence similarity and feature extraction tasks. As part of the Infinity project, it serves as the stable default model for generating high-quality text embeddings while maintaining a relatively small parameter footprint of 33.4M parameters.
Implementation Details
The model can be deployed using the infinity_emb package, offering flexible deployment options including GPU acceleration with PyTorch and CPU optimization through ONNX. It supports both synchronous and asynchronous embedding generation, with built-in support for flash attention on GPU implementations.
- Supports both PyTorch and ONNX inference engines
- Compatible with CPU and CUDA devices
- Implements flash attention for GPU optimization
- Offers torch.compile support for enhanced performance
Core Capabilities
- Sentence embedding generation
- Feature extraction for NLP tasks
- Sentence similarity computation
- Efficient processing through multiple inference backends
Frequently Asked Questions
Q: What makes this model unique?
The model stands out for its efficient architecture, balancing performance with a compact size of 33.4M parameters, while providing robust embedding capabilities through the Infinity framework.
Q: What are the recommended use cases?
The model is ideal for applications requiring sentence embeddings, text similarity comparisons, and feature extraction tasks. It's particularly well-suited for production environments where both CPU and GPU deployment options are needed.