run.md

Data Preparation

Prepare datasets in data directory. You can reference each datasets format on each projects README.

clshub/data
└── rsna2022

Environment setup

Clone repo

$ git clone https://github.com/okotaku/clshub

Start a docker container

$ docker compose up -d clshub

Prepare configs

For basic usage of configs, see MMClassification: Learn about Configs

Train a model

# single-gpu
$ docker compose exec clshub mim train mmcls ${CONFIG_FILE}
# Example
$ docker compose exec clshub mim train mmcls configs/projects/rsna2022/efficientnet/efficientnet-b3_2xb8_rsna2022.py

# multiple-gpu
$ docker compose exec clshub mim train mmcls ${CONFIG_FILE} --gpus ${GPUS} --launcher pytorch

Test a dataset

# single-gpu
$ docker compose exec clshub mim test mmcls ${CONFIG_FILE} --checkpoint ${CHECKPOINT_FILE}
# Example
$ docker compose exec clshub mim test mmcls configs/projects/rsna2022/efficientnet/efficientnet-b3_2xb8_rsna2022.py --checkpoint work_dirs/efficientnet-b3_2xb8_rsna2022/best_rsna2022/f1-score_cancer_epoch_15.pth

More details

See MMClassification Docs