Tengine is a lite, high performance, modular inference engine for embedded device

fastllm是后端无依赖的高性能大模型推理库。同时支持张量并行推理稠密模型和混合模式推理MOE模型，任意10G以上显卡即可推理满血DeepSeek。双路9004/9005服务器+单显卡部署DeepSeek满血满精度原版模型，单并发20tps；INT4量化模型单并发30tps，多并发可达60+。

Use TensorRT API to implement Caffe-SSD， SSD（channel pruning）， Mobilenet-SSD

Caffe implementation of FAIR paper "Focal Loss for Dense Object Detection" for SSD.

GraspSplats: Efficient Manipulation with 3D Feature Splatting