mackelab / delfi

Density estimation likelihood-free inference. No longer actively developed see https://github.com/mackelab/sbi instead
http://www.mackelab.org/delfi
Other
71 stars 28 forks source link

Gabor RF project updates #43

Closed mnonnenm closed 5 years ago

mnonnenm commented 5 years ago

merging development forks

fixes and features for SNPE-A (as used in Gabor RF subproject):

* backup-predict (uncorrected for proposal) if returned precision matrices non-p.s.d. 
* numerically stable predict, numerically stable predict from MoG-proposals
* new code for splitting components in last round (based on standardized MoG initialization written in BaseDistribution originally for SNPE-B);
* support for Simulation-based Calibration  (SNPE-A can meaningfully track calibration against the proposal prior even in later rounds)

new CNN-related features (as used in Gabor RF subproject):

* improved control over CNN architecture (strides & pooling layers,  shapes of input-nomralization parameters)

* extra_inputs: features bypass initial convolutional (or recurrent) layers and directly feed into fully connected layers
jan-matthis commented 5 years ago

Could you remove the changes to .travis.yml and test_inference.py from the PR?