First build the image:
docker build \
--target production \
-t adval:latest \
.
and then run the container:
docker run \
--init \
-p 8888:8888 \
--mount type=bind,source=.,target=/app \
adval:latest
Jupyter notebooks will be available for review & reproduction at: http://127.0.0.1:8888/tree (requires token):
- demo-house-prices.ipynb: where train & test datasets follow the same underlying distribution.
- demo-sberbank-russian-housing-market.ipynb: where train & test datasets follow different underlying distributions.