This repository is not an original official implementation of the work, but a refactored codebase. Performed within the FSE course at Skoltech.
Follow this quick start instructions step by step in order to download, build and run code inside docker containers:
make hub_download
- download datamake hub_prepare
- prepare datamake hub_train
- train netmake hub_eval
- evaluate netmake hub_lint
- lint all.py
files in repo
- In order to change run configuration of prepare, train and test, you should change the following flags inside Makefile:
PREPARE_FLAGS
,TRAIN_FLAGS
andTEST_FLAGS
. DATA_DIR
variable in Makefile defines the directory that is used for storing the data.- In order to use additional libraries during train, test and steps, you should add them to the variable
REQ
insetup.py
files in the repositoriesdocker/train
,docker/eval
anddocker/prepare
. - If you want to add tests of the particular functionality of
prepare.py / train.py / test.py
you need to add your test cases inunit_test.py
file that correspons to the.py
file whom functionality you would like to check. - Note that in current solution tests run simultaneously with the docker container running.