boostcamp_pstage10

Environment

1. Docker

docker run -it --gpus all --ipc=host -v ${path_to_code}:/opt/ml/code -v ${path_to_dataset}:/opt/ml/data placidus36/pstage4_lightweight:v0.4 /bin/bash

2. Install dependencies

pip install -r requirements.txt

Run

1. train

python train.py --model_config ${path_to_model_config} --data_config ${path_to_data_config}

2. inference(submission.csv)

python inference.py --model_config configs/model/mobilenetv3.yaml --weight exp/2021-05-13_16-41-57/best.pt --img_root /opt/ml/data/test --data_config configs/data/taco.yaml3

Reference

Our basic structure is based on Kindle(by JeiKeiLim)