Welcome to mirror list, hosted at ThFree Co, Russian Federation.

github.com/clementfarabet/lua---nnx.git - Unnamed repository; edit this file 'description' to name the repository.
summaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2011-09-21fixing countersMarco Scoffier
2011-09-21forgot to copy params from CPU to GPUMarco Scoffier
2011-09-21added some (ugly) code for optimizing on CPU when evaluating function on GPUMarco Scoffier
2011-09-21Fixed MAJOR bug. The new backward() doesnt update weights anymore.Clement Farabet
So any module should call its accGradParameters().
2011-09-20Fixed BatchOptimization in // mode.Clement Farabet
2011-09-20Removed redundant instantiations.Clement Farabet
2011-09-20Fixed (?) flattenParameters.Clement Farabet
2011-09-16Fixed typoClement Farabet
2011-09-16Added pre-hook code in BatchOptimization.Clement Farabet
2011-09-16Merge branch 'master' of github.com:clementfarabet/lua---nnxClement Farabet
2011-09-16Rewrote spatial upsampler to use raw pointers.Clement Farabet
2011-09-16if/else scope bugMarco Scoffier
2011-09-16talk with Clement better integrationMarco Scoffier
2011-09-15changes to run notMNIST with paddingMarco Scoffier
2011-09-15OupsClement Farabet
2011-09-14Got rid of SpatialMaxPoolingClement Farabet
2011-09-14SuperCriterion accepts multiple targets.Clement Farabet
2011-09-14Added a flag to optionally bypass the creation of a spatial output.Clement Farabet
2011-09-13Adding SpatialClassifier (forgot it from previous commit)Clement Farabet
2011-09-13Merge branch 'master' of github.com:clementfarabet/lua---nnxClement Farabet
2011-09-13Introducing: SpatialClassifier, to replace all Spatial modules.Clement Farabet
The goal is simple: instead of writing tons of Spatial*** modules, we make sure all modules (Linear,SoftMax,...) can handle batches, and use SpatialClassifier to replicate these modules spatially.
2011-09-12Added support for CUDA tensor in DataList.Clement Farabet
2011-09-12ConfusionMatri is now agnostic to type (cuda compat)Clement Farabet
2011-09-10updated DataSet to accept png and not automatic rescale. defaults are ↵Marco Scoffier
identical to before
2011-09-10Default for lBFGS is not to conserve its state across mini-batches.Clement Farabet
2011-09-10Using cleaner function in batchoptimization.Clement Farabet
2011-09-08Reincluded Sparse Conv.Clement Farabet
2011-09-08Added Sparse Conv for backward compat.Clement Farabet
2011-09-07Updated training code to latest Torch version: backward is now split into two.Clement Farabet
2011-09-06Using new gnuplot.Clement Farabet
2011-09-06Mini batch is now distributed to workers according to their speed !!Clement Farabet
2011-09-05Switch close/error.Clement Farabet
2011-09-05Actual error when something goes wrong in mapreduce setup.Clement Farabet
2011-09-05Protected map-reduce setup.Clement Farabet
2011-09-05Added serializing methods for confusion matrix.Clement Farabet
2011-09-04Using new sfork function.Clement Farabet
2011-09-04Fixed stupid bug for SGD.Clement Farabet
2011-09-04Re-thought map-reduce.Clement Farabet
2011-09-04Fixed typoeClement Farabet
2011-09-04BatchOptimization now supports server list.Clement Farabet
2011-09-04using new parallel API: fork/exec.Clement Farabet
2011-09-02Fixed tester.Clement Farabet
2011-09-02Formatting.Clement Farabet
2011-09-02Fixed tabs.Clement Farabet
2011-09-02make parallel compatible with not optimize hookMarco Scoffier
2011-09-01added flag to quit after a specified number of function evaluationsMarco Scoffier
2011-09-01bug in batch creationMarco Scoffier
2011-09-01added batchTest and hookMarco Scoffier
2011-09-01Re-organized BatchOptimization modules.Clement Farabet
2011-08-31Merge branch 'master' of github.com:clementfarabet/lua---nnxClement Farabet
generated by cgit v1.2.3 (git 2.25.1) at 2024-08-03 04:59:17 +0300