TensorRT-使用TensorRT部署3D-Resnet视频分类算法-优质算法部署项目实战.zip

# Activity Recognition TensorRT Perform video classification using 3D ResNets trained on Kinetics-700 and Moments in Time dataset - accelerated with TensorRT 8.0 [] ## TensorRT 8 Installation Assuming you have CUDA already installed, go ahead and download TensorRT 8 from [here](https://developer.nvidia.com/tensorrt-getting-started). Follow instructions of installing the system binaries and python package for tensorrt [here](https://docs.nvidia.com/deeplearning/tensorrt/install-guide/index.html#installing-tar). ## Python dependencies Install the necessary python dependencies by running the following command - ```sh pip3 install -r requirements.txt ``` ## Clone the repository This is a straightforward step, however, if you are new to git recommend glancing threw the steps. First, install git ```sh sudo apt install git ``` Next, clone the repository ## Download Pretrained Models Download models from [google-drive](https://drive.google.com/drive/folders/1PumnUl_-eVvk0tFpn463vPxpqchGEuWB?usp=sharing) and place them in the current directory. ## Running the code The code supports a number of command line arguments. Use help to see all supported arguments ```sh ➜ python3 action_recognition_tensorrt.py --help usage: action_recognition_tensorrt.py [-h] [--stream STREAM] [--model MODEL] [--fp16] [--frameskip FRAMESKIP] [--save_output SAVE_OUTPUT] Action Recognition using TensorRT 8 optional arguments: -h, --help show this help message and exit --stream STREAM Path to use video stream --model MODEL Path to model to use --fp16 To enable fp16 precision --frameskip FRAMESKIP Number of frames to skip --save_output SAVE_OUTPUT Save output as video ``` Run the script this way: ```sh # Video python3 action_recognition_tensorrt.py --stream /path/to/video --model resnet-18-kinetics-moments.onnx --fp16 --frameskip 2 # Webcam python3 action_recognition_tensorrt.py --stream webcam --model resnet-18-kinetics-moments.onnx --fp16 --frameskip 2 ```