fizyr/keras-retinanet 0.3
Gaiser, Hans, Maarten de Vries, Williamson, Ashley, Valeriu Lacatusu, , András, Yann Henon, jjiun, Cristian Gratie, Morariu, Mihai, Enrico Liscio, Rodrigo Meira de Andrade, Conceição, Pedro, hannesedvartsen, Wudi Fang, Clark, Mike, Iver Jordal, Max Van Sande, Ori, Pierre Dérian, Muhammed Kocabas · Zenodo (CERN European Organization for Nuclear Research) · 2018
Changes since last release Implement DenseNet, VGG backbones. Add option to freeze backbone layers. Add logging of evaluation to tensorboard. Add pretty colors for 80 classes. Fix batch_size > 1 issues. Refactor model outputs (should hopefully stay like this now). Simplified training by splitting into "training model" and "inference model". Add structure for backbone specific functions (such as load_model). Encode regression as x1/y1/x2/y2 offsets (increases mAP to 0.350, previously 0.345). Use nearest upsampling method. Credits to: @vidosits @cgratie @DiegoAgher @eduramiba @GuillaumeErhard @Muhannes @hgaiser @iver56 @jjiunlin @srslynow @de-vri-es @Ori226 @pedroconceicao @pderian @rodrigo2019 @lvaleriu @yhenon