https://github.com/GPflow/GPflow
Revision 77be468c95e06730f613726ec7a09b9594602b46 authored by John Bradshaw on 10 October 2017, 10:53:08 UTC, committed by John Bradshaw on 10 October 2017, 10:53:08 UTC
* fixed a bug in log likelihood that was causing NaNs * fixed the gradients for likelihood as was doing too many tf.stop_gradients * added the option to not rerun EP each time you call log_likelihood * tidied up the demo so that get nice clean comparison between VGP and EP * soem numerical fixes and minor refactoring of EP code.
1 parent f7b52ab
Tip revision: 77be468c95e06730f613726ec7a09b9594602b46 authored by John Bradshaw on 10 October 2017, 10:53:08 UTC
Binary classification with EP
Binary classification with EP
Tip revision: 77be468
run_tests.sh
#!/bin/bash
# Script for running GPflow tests in sequential and parallel modes.
# Running tensorflow based tests in distinct processes prevents
# bad memory accumulations which can lead to crashes or slow runs
# on resource limited hardware.
# Written by Artem Artemev, 06/08/2017
set -e
mode=${1:-"--sequential"}
case "$mode" in
-p|--parallel)
numproc=$([[ $(uname) == 'Darwin' ]] && sysctl -n hw.physicalcpu_max || nproc)
echo ">>> Parallel mode. Number of processes = $numproc"
echo testing/test_*.py | xargs -n 1 -P "$numproc" bash -c 'nosetests -v --nologcapture $0 || exit 255'
;;
-s|--sequential)
for test_file in testing/test_*.py; do
echo ">>> Run $test_file"
nosetests -v --nologcapture "$test_file"
rc=$?
if [ "$rc" != "0" ]; then
echo ">>> $test_file failed"
exit $rc
fi
done
;;
*)
;;
esac
Computing file changes ...