GPU memory variation for the influence of the co-training-inference
Jetson TX2 8GB version
jetson-inference -> https://github.com/dusty-nv/jetson-inference.git
jeston-stats
sudo -H pip3 install jeston-stats
torch==1.7.0 torchvision==0.8.1
train.py -----> training job wih pytorch framework
log_gpu.py -----> gpu monitoring program
inference.py -----> inference task with TensorRT engine
/data -----> training dataset dir
-----/class1
------------/train
------------/val
------------/test
-----/class2
------------/train
------------/val
------------/test
start_validation.sh -----> variation procedure execution script
generate_dataset.py
demo.jpg
- Finish the system setting including installation above dependencies.
- Run start_validation.sh
- Get the GPU memory variation from log.txt