https://github.com/GPflow/GPflow
Tip revision: 5c6924058fe8b5fe2cca25aaa8427d12951c89b4 authored by hrs13 on 29 January 2019, 16:10:40 UTC
changed behaviour of multisample kernels. Now cov(SND,SMD) assumes that the samples are paired over S and returns SNM (rather than SNSM), and cov(ND,SMD) is SNM (rather than NSM)
changed behaviour of multisample kernels. Now cov(SND,SMD) assumes that the samples are paired over S and returns SNM (rather than SNSM), and cov(ND,SMD) is SNM (rather than NSM)
Tip revision: 5c69240
config.yml
runtest: &runtest
steps:
- checkout
- run:
name: Install GPflow
command: python setup.py install
- run:
name: Run tests
command: pytest -v -W ignore::UserWarning --durations=10 -m "${PYTEST_FILTER}" --cov=./gpflow ./tests
- run:
name: Upload coverage report
command: bash <(curl -s https://codecov.io/bash) -t "${CODECOV_TOKEN}"
version: 2.1
jobs:
unit-test:
<<: *runtest
docker:
- image: awav/tensorflow:1.12.0
environment:
PYTEST_FILTER: not notebooks
notebook-test:
<<: *runtest
docker:
- image: awav/tensorflow:1.12.0
environment:
PYTEST_FILTER: notebooks
workflows:
version: 2.1
build_and_test:
jobs:
- unit-test
- notebook-test