3
97%
master: 97%

Ran 26 Aug 2016 02:28PM UTC

Files 45

Run time 59s

Badge

Embed ▾

Committed 26 Aug 2016 02:25PM UTC coverage: 89.791% (+0.01%) from 89.779%

Job # PYTHON_VERSION="3.5" COVERAGE="true" LATEST="true" "DEAP_VERSION=1.0.1" "XGBOOST_VERSION=0.4a30"

Build Type

push

travis-ci

Committed by

rhiever

Commit Message

Bugfix: random_state was set only on init

Previously, TPOT only set the `random_state` (i.e., RNG seed) on init.
This led to unintended behavior when TPOT was used in sklearn methods
such as `cross_val_score`, which uses `set_params` to set parameters
such as the `random_state` instead of creating a new TPOT instance.

As such, I made the change to store `random_state` internally on init
then actually set the seed when `fit` is called. Now if we use
`cross_val_score`, the same random state is provided at the beginning
of each TPOT run.

Run Details

774 of 862 relevant lines covered (89.79%)

0.9 hits per line

rhiever / tpot / 548 / 3
97%
master: 97%

README BADGES
x

Markdown

Textile

RDoc

HTML

Rst

Source Files on job 548.3 (PYTHON_VERSION="3.5" COVERAGE="true" LATEST="true" "DEAP_VERSION=1.0.1" "XGBOOST_VERSION=0.4a30")

rhiever / tpot / 548 / 3 97% master: 97%

README BADGES x

Markdown

Textile

RDoc

HTML

Rst

Source Files on job 548.3 (PYTHON_VERSION="3.5" COVERAGE="true" LATEST="true" "DEAP_VERSION=1.0.1" "XGBOOST_VERSION=0.4a30")

rhiever / tpot / 548 / 3
97%
master: 97%

README BADGES
x