This repository contains model designs, deployment solutions for state-of-the-art networks, and inference samples for Autonomous Vehicle (AV) applications on NVIDIA DRIVE Platforms.

• Deployment and Inference Solutions

  • ONNX Export Guidance for TensorRT
  • BEVFormer INT8 Explicit Quantization
  • DCNv4 TensorRT
  • Far3D TensorRT
  • Deploy LLMs with TensorRT-LLM
  • MTMI TensorRT
  • PETRv1&v2 TensorRT
  • Sparsity INT8 Training and TensorRT Inference
  • StreamPETR TensorRT
  • UniAD TensorRT
  • VAD-TensorRT

• Hardware-friendly Models

  • DEST
  • ReduceFormer
  • Swin-Free