Commits

Breadcrumbs

History for

onmaster

User selector

All users

All time

Commit History

Commits on Oct 31, 2015

• Merge pull request #3082 from gustavla/pycaffe-snapshot

  Show description for f5fd18b
  shelhamer

  committed
  f5fd18b

  Copy full SHA for f5fd18b

Commits on Oct 17, 2015

• Change solver type to string and provide solver registry

  ronghanghu

  committed
  0eea815

  Copy full SHA for 0eea815

• Split solver code into one file per solver class

  ronghanghu

  committed
  b822a70

  Copy full SHA for b822a70

Commits on Sep 18, 2015

• refine format of switch case in solver

  yanchen036

  committed
  f75d594

  Copy full SHA for f75d594

• Expose `Snapshot` to pycaffe

  Show description for b4f9add
  gustavla

  committed
  b4f9add

  Copy full SHA for b4f9add

Commits on Sep 14, 2015

• Check that the snapshot directory is writeable before starting training

  

  Sean Bell

  committed
  ab554cb

  Copy full SHA for ab554cb

Commits on Aug 22, 2015

• Add signal handler and early exit/snapshot to Solver.

  Show description for ff19d5f
  jyegerlehner

  committed
  ff19d5f

  Copy full SHA for ff19d5f

Commits on Aug 14, 2015

• Cite Adam paper in solver.hpp

  ronghanghu

  committed
  bf42e6e

  Copy full SHA for bf42e6e

• Adam solver

  Show description for 4e4c89b
  

  PatWie

  authored and

  ronghanghu

  committed
  4e4c89b

  Copy full SHA for 4e4c89b

Commits on Aug 13, 2015

• Merge pull request #2903 from ronghanghu/multi_gpu

  Show description for bb0a90e
  ronghanghu

  committed
  bb0a90e

  Copy full SHA for bb0a90e

Commits on Aug 12, 2015

• Data Layers Parallel for Multi-GPU

  Show description for 0d34d5b
  ronghanghu

  committed
  0d34d5b

  Copy full SHA for 0d34d5b

Commits on Aug 10, 2015

• Clean up and modernize AdaDelta code; add learning rate support; add additional test cases

  matthiasplappert

  committed
  f2e523e

  Copy full SHA for f2e523e

• Updated AdaDelta for modern Caffe; reduced iterations on multi-iter tests

  

  kevinbache

  authored and

  matthiasplappert

  committed
  4c58741

  Copy full SHA for 4c58741

• Implement AdaDelta; add test cases; add mnist examples

  

  mohomran

  authored and

  matthiasplappert

  committed
  1ce3380

  Copy full SHA for 1ce3380

Commits on Aug 9, 2015

• Multi-GPU

  Show description for e5575cf
  

  cypof

  authored and

  shelhamer

  committed
  e5575cf

  Copy full SHA for e5575cf

• Implement RMSProp Solver

  Show description for abe99e8
  

  erogol

  authored and

  ronghanghu

  committed
  abe99e8

  Copy full SHA for abe99e8

Commits on Aug 7, 2015

• Snapshot model weights/solver state to HDF5 files.

  Show description for ada055b
  erictzeng

  committed
  ada055b

  Copy full SHA for ada055b

Commits on May 28, 2015

• directly normalize accumulated gradients

  Show description for 0e7a078
  shelhamer

  committed
  0e7a078

  Copy full SHA for 0e7a078

Commits on May 27, 2015

• Solver::MakeUpdate() -> Solver::ApplyUpdate

  Show description for 76db47e
  shelhamer

  committed
  76db47e

  Copy full SHA for 76db47e

• Refactor solvers regularization and logging code

  

  cypof

  authored and

  shelhamer

  committed
  ca81667

  Copy full SHA for ca81667

Commits on Mar 5, 2015

• expose Solver::Restore() as public and Solver.restore() in pycaffe

  Show description for c0219cc
  shelhamer

  committed
  c0219cc

  Copy full SHA for c0219cc

Commits on Feb 14, 2015

• Add gradient clipping -- limit L2 norm of parameter gradients

  jeffdonahue

  committed
  f38ddef

  Copy full SHA for f38ddef

Commits on Jan 3, 2015

• break out Step from Solver

  longjon

  committed
  55d7ffd

  Copy full SHA for 55d7ffd

Commits on Oct 10, 2014

• Added Multistep, Poly and Sigmoid learning rate decay policies

  Show description for b025da7
  sguada

  committed
  b025da7

  Copy full SHA for b025da7

Commits on Oct 1, 2014

• add accessor for Solver::iter_

  Show description for e222c57
  longjon

  committed
  e222c57

  Copy full SHA for e222c57

Commits on Sep 3, 2014

• solver.hpp: add \briefs

  

  jeffdonahue

  authored and

  shelhamer

  committed
  4572663

  Copy full SHA for 4572663

Commits on Sep 1, 2014

• lint

  

  qipeng

  authored and

  jeffdonahue

  committed
  06f335f

  Copy full SHA for 06f335f

• Added sanity check for AdaGradSolver; added MNIST examples for solvers

  

  qipeng

  authored and

  jeffdonahue

  committed
  c1ff97c

  Copy full SHA for c1ff97c

• Merge Test{SGD,AdaGrad,Nesterov}Solver; they become subclasses of

  Show description for f206c64
  jeffdonahue

  committed
  f206c64

  Copy full SHA for f206c64

• Addressed Yangqing's comments

  

  qipeng

  authored and

  jeffdonahue

  committed
  3f7a910

  Copy full SHA for 3f7a910

• Added L1 regularization support for the weights

  

  qipeng

  authored and

  jeffdonahue

  committed
  a683c40

  Copy full SHA for a683c40

• converted pointers to shared_ptr

  

  qipeng

  authored and

  jeffdonahue

  committed
  ed8b1da

  Copy full SHA for ed8b1da

• restored vituals in solver.hpp

  

  qipeng

  authored and

  jeffdonahue

  committed
  8a9c268

  Copy full SHA for 8a9c268

• Solver switching support & implementation of Nesterov's accelerated gradient and AdaGrad

  

  qipeng

  authored and

  jeffdonahue

  committed
  9d10569

  Copy full SHA for 9d10569

Commits on Aug 13, 2014

• Store loss coefficients in layer; use for prettier training output.

  jeffdonahue

  committed
  415123b

  Copy full SHA for 415123b