Welcome to
mirror list
, hosted at
ThFree Co
, Russian Federation.
github.com/torch/optim.git - Unnamed repository; edit this file 'description' to name the repository.
index
:
github.com/torch/optim.git
assertions
conffix
master
nanfix
require
revert-113-sgd-lrs-fix
revert-63-confusion_matrix_fix
split
Unnamed repository; edit this file 'description' to name the repository.
www-data
summary
refs
log
tree
commit
diff
log msg
author
committer
range
path:
root
/
sgd.lua
Age
Commit message (
Expand
)
Author
2016-06-30
Fix bad alignment, trailing spaces and tabs
Alfredo Canziani
2016-06-03
Revert "Fix bug with sgd individual learning rates"
revert-113-sgd-lrs-fix
Soumith Chintala
2016-06-02
Fix bug with individual learning rates
Geoff Pleiss
2015-11-13
documentation update for config.learningRates, state.evalCounter
Arulkumar
2014-11-30
copy variables into state params for type casting
Aysegul Dundar
2014-11-28
user either sets wd or wds
Aysegul Dundar
2014-11-24
more efficient way of state parameters update
Aysegul Dundar
2014-11-24
weight decay for individual param option added
Aysegul Dundar
2013-12-05
Better formatting for docstrings in REPL
Andreas Fidjeland
2013-12-03
Turned documentation into markdown format
Andreas Fidjeland
2013-08-19
Allowed to enable Nesterov momentum.
Ivo Danihelka
2013-05-24
New dampening term.
Clement Farabet
2013-05-21
CRITICAL: weight decay is finally properly implemented!
Clement Farabet
2013-02-19
add a 4th parameter to interface to split configuration parameters
koray kavukcuoglu
2012-09-28
Updated info / help.
Clement Farabet
2012-01-22
Merge branch 'lbfgs'
Clement Farabet
2012-01-22
All debugged L-BFGS implementation.
Clement Farabet
2012-01-19
Fixed weight decay to scale with learning rate.
Clement Farabet
2012-01-18
Fixed bugs in SGD and ASGD implementations.
Clement Farabet
2011-11-03
Added asgd.
Clement Farabet
2011-11-03
Fixed eval counter for SGFixed eval counter for SGDD
Clement Farabet
2011-11-03
Added first SGD, and orgniazed package.
Clement Farabet