facebookarchive/caffe2

MKL-DNN integration does not support backward pass

Open

#2.376 geöffnet am 22. März 2018

Auf GitHub ansehen
 (1 Kommentar) (0 Reaktionen) (0 zugewiesene Personen)Shell (8.424 Stars) (1.954 Forks)batch import
help wanted

Beschreibung

MKL-DNN integration (in caffe2/mkl) implements only forward pass operators, although MKL-DNN itself provides primitives for both forward and backward pass. Implementation of ConvGradient operator would let Caffe2 use MKL-DNN for training on CPU, and support some models which need backprop (e.g. deep dream).

Contributor Guide