Skip to content

README.md

Latest commit

 

History

History
62 lines (51 loc) · 1.49 KB

README.md

File metadata and controls

62 lines (51 loc) · 1.49 KB

NeuFlow

How to Run (Pytorch)

  1. set up a virtual environment.
cd NeuFlow
git clone https://github.com/neufieldrobotics/NeuFlow_v2.git
cd NeuFlow_v2

# Create a new conda environment with Python 3.11
conda create -n neuflow -y python=3.11

# Activate the created environment
conda activate neuflow

# Install the required Python packages
pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu128
pip install opencv-python
pip install huggingface-hub
  1. run the original pytorch model on test images.
python infer.py
  1. check pytorch model inference performance
cd ..
python ../gen_video2imgs.py
python infer.py
  • input size: 288 x 512, iters:
  • 249 iterations time: 11.3614 [sec]
  • Average FPS: 21.92 [fps]
  • Average inference time: 45.63 [msec]

How to Run (TensorRT)

  1. generate onnx file
python onnx_export.py
// a file '.onnx' will be generated in onnx directory.
  1. build tensorrt model and run
conda activate trte
python onnx2trt.py
// a file '.engine' will be generated in engine directory.
  • input size: 288 x 512, iters:
  • 249 iterations time: 9.6196 [sec]
  • Average FPS: 25.88 [fps]
  • Average inference time: 38.63 [msec]

Back