Docker - Yolov7 - NVIDIA (Triton-Server/Deepstream)

This repository provides an out-of-the-box deployment solution for creating an end-to-end procedure to train, deploy, and use Yolov7 models on Nvidia GPUs using Triton Server and Deepstream.

Training and Quantize Yolov7

Use yolov7 folder
This Repo sets up an environment for running NVIDIA PyTorch applications, focusing on training YOLOv7 models, including quantization and profiling for achieving optimal performance with minimal accuracy loss. It deploys YOLOv7 with YOLO Quantization-Aware Training (QAT) patched. It also installs the TensorRT Engine Explorer (TREx), which is a Python library and a set of Jupyter notebooks for exploring a TensorRT engine plan and its associated inference profiling data.

Deploy Yolov7 Models on NVIDIA Triton-Server

Use triton-server-yolov7 folder
Docker Image to Build Yolov7 models on Triton-Server This repository serves as an example of deploying the Models YOLOv7 model (FP16) and the YOLOv7 QAT (INT8) on Triton-Server for performance and testing. It includes support for applications developed using Nvidia DeepStream. Instructions to deploy YOLOv7 as TensorRT engine to Triton Inference Server.

Deploy NVIDIA Deepstream and Sample App

Use deepstream-yolov7 folder
This repo provides a set of instructions for building a Docker image tailored for deploying a Sample Deepstream application with support for YOLOv7 model inference served by Triton Server.