the 9th place solution code of Kaggle RSNA STR Pulmonary Embolism Detection (https://www.kaggle.com/c/rsna-str-pulmonary-embolism-detection/overview)
Please read this post on Kaggle Discussion.
Our team mainly use cloud machine with following spec on GCP.
OS: Ubuntu 18.04 LTS
GPU: Tesla V100 x4
You should install docker-ce. If you don't have docker installed, please refer to this page to install it.
Plese download data to ./input
from https://www.kaggle.com/c/rsna-str-pulmonary-embolism-detection/data and unzip it.
$ docker build ./images -t kaggle/pytorch:rsna
$ sh ./bin/preprocess.sh
# Train and extract feature
$ sh ./bin/stage_1_train.sh
$ sh ./bin/stage_1_predict_valid.sh
# Extract feature from test data
$ sh ./bin/stage_1_predict_test.sh
# Train
$ sh ./bin/stage_2_384_train.sh
$ sh ./bin/stage_2_512_train.sh
$ sh ./bin/stage_2_concat_train.sh
# Predict test data
$ sh ./bin/stage_2_384_test.sh
$ sh ./bin/stage_2_512_test.sh
$ sh ./bin/stage_2_concat_test.sh
# Train
$ sh ./bin/stacking_train.sh
# Predict test data
$ sh ./bin/stacking_test.sh
$ sh ./bin/blending.sh
The final output is generated in ./output/final_submission.csv
.