usls¶
A cross-platform Rust library powered by ONNX Runtime for efficient inference of SOTA vision and vision-language models.

✨ Features¶
-
High Performance
Multi-threading, SIMD, and CUDA-accelerated processing with ONNX Runtime execution providers
-
Cross-Platform
Linux, macOS, Windows with support for CUDA, TensorRT, CoreML, OpenVINO, DirectML
-
Unified API
Single
Modeltrait withrun()/forward()and unifiedYoutput for all models -
Auto-Management
Automatic model download from HuggingFace/GitHub, caching and path resolution
-
Multiple Inputs
Support for images, directories, videos, webcam, streams and combinations
-
Precision Support
FP32, FP16, INT8, UINT8, Q4, Q4F16, BNB4, and more quantization options
-
Full-Stack Suite
Complete workflows with
DataLoader,Annotator, andViewer -
Model Ecosystem
50+ SOTA vision and VLM models ready to use