8000 Loss generalization by jeffdonahue · Pull Request #686 · BVLC/caffe · GitHub

More Web Proxy on the site http://driver.im/

Loss generalization#686

Merged

jeffdonahue merged 10 commits intoBVLC:devfrom

jeffdonahue:loss-generalization

Aug 13, 2014

Commits on Aug 13, 2014

Add loss_weight to proto, specifying coefficients for each top blob
jeffdonahue
committed
Add net tests for loss_weight.
jeffdonahue
committed
Generalize loss by allowing any top blob to be used as a loss in which
jeffdonahue
committed
Make multiple losses work by inserting split layers and add some tests for it.
jeffdonahue
committed
Net::Init can determine that layers don't need backward if they are not
jeffdonahue
committed
AccuracyLayer only dies when necessary.
jeffdonahue
committed
Disallow in-place computation in SPLIT layer -- has strange effects in
jeffdonahue
committed
Also display outputs in the train net. (Otherwise, why have them?)
jeffdonahue
committed
Add ACCURACY layer and softmax_error output to lenet_consolidated_solver
jeffdonahue
committed
Store loss coefficients in layer; use for prettier training output.
jeffdonahue
committed

0