pymc-devs / pymc3 / 8682 / 3
87%
master: 87%

DEFAULT BRANCH: master
Ran
Files 141
Run time 13s
Badge
Embed ▾
README BADGES
x

If you need to use a raster PNG badge, change the '.svg' to '.png' in the link

Markdown

Textile

RDoc

HTML

Rst
coverage: 25.737% (-0.2%) from 25.948%
FLOATX='float32' TESTCMD="--durations=10 --cov-append pymc3/tests/test_examples.py pymc3/tests/test_posteriors.py"

push

travis-ci

junpenglao 
Mixture random cleanup (#3364)

* WIP mixture

* Still WIP but much better performance on posterior predictive sampling. Still need to work around size issues.

* Fixed mixture random method (and also multinomial when using a Cholesky covariance matrix).

* Fixed float32 all close relative tolerance error.

* Changed the way in which we wrap the comp_dists.random method to something less hackish. At least we dont ignore lint anymore.

* Moved comp_dists type checking to __init__. Changed _comp_logp, mean and median to stack on the last axis (which should allow Mixture to work with multidimensional distributions). Changed docstring a bit. Minor change in _comp_samples to enfore dtype.

* Improved support for Mixtures of multidimensional distributions.

* Added tests for mixtures of multidimensional distributions.

* Added kwarg to generate_samples to help pass raw_size_ tuples

* Cleaned up infer_comp_dist_shapes and added docstring.

* Fixed latent component assignment for mixtures of multidimensional distributions.

* Fixed mixture of mixtures errors. The error was in broadcast_distribution_samples.

* Patched broadcast_distribution_samples. Must do a cleaner fix.

* Fixed broadcast_distribution_samples bugs. Added release notes.

* Added tests comparing mixtures to latent component models

* Added more examples to Mixture docstring.

5296 of 20577 relevant lines covered (25.74%)

0.26 hits per line

Source Files on job 8682.3 (FLOATX='float32' TESTCMD="--durations=10 --cov-append pymc3/tests/test_examples.py pymc3/tests/test_posteriors.py")