Commits

Breadcrumbs

History for

onmaster

User selector

All users

All time

Commit History

Commits on Dec 28, 2015

• add support for 2D dilated convolution

  

  fyu

  authored and

  longjon

  committed
  93bfcb5

  Copy full SHA for 93bfcb5

Commits on Dec 2, 2015

• dismantle layer headers

  Show description for 300f43f
  shelhamer

  committed
  300f43f

  Copy full SHA for 300f43f

Commits on Oct 20, 2015

• Clean redundant/unnecessary headers

  kkhoot

  committed
  80d0452

  Copy full SHA for 80d0452

Commits on Sep 20, 2015

• fix broken conv/deconv reshaping caused by reading bottom shape in

  Show description for 6a00eca
  longjon

  committed
  6a00eca

  Copy full SHA for 6a00eca

Commits on Sep 19, 2015

• Im2col and Convolution layers support N spatial axes

  jeffdonahue

  committed
  9d8206e

  Copy full SHA for 9d8206e

Commits on May 27, 2015

• accumulate gradients in (de)conv layers

  

  longjon

  authored and

  shelhamer

  committed
  8cc9af0

  Copy full SHA for 8cc9af0

Commits on Feb 5, 2015

• Layer type is a string

  jeffdonahue

  committed
  3b13846

  Copy full SHA for 3b13846

Commits on Jan 11, 2015

• rewrite ConvolutionLayer to use BaseConvolutionLayer helpers

  longjon

  committed
  e3e2f2d

  Copy full SHA for e3e2f2d

Commits on Dec 1, 2014

• remove redundant code in ConvolutionLayer::Reshape

  longjon

  committed
  8009436

  Copy full SHA for 8009436

Commits on Oct 13, 2014

• move registration code to corresponding cpp files.

  

  Yangqing Jia

  committed
  2f88dab

  Copy full SHA for 2f88dab

Commits on Sep 20, 2014

• combine col_{data,diff} into single col_buff to halve memory usage

  Show description for e8ba4c8
  shelhamer

  committed
  e8ba4c8

  Copy full SHA for e8ba4c8

• optimize 1x1 convolution for Network-in-Network style layers

  Show description for 8109a6e
  shelhamer

  committed
  8109a6e

  Copy full SHA for 8109a6e

Commits on Sep 19, 2014

• fix types of (Layer)SetUp, Reshape, Forward, and Backward calls

  Show description for 31326a1
  longjon

  committed
  31326a1

  Copy full SHA for 31326a1

Commits on Sep 18, 2014

• split off Reshape for vision layers

  Show description for 6c63b8c
  

  longjon

  authored and

  shelhamer

  committed
  6c63b8c

  Copy full SHA for 6c63b8c

• [docs] comment ConvolutionLayer

  shelhamer

  committed
  18ca362

  Copy full SHA for 18ca362

Commits on Sep 8, 2014

• Merge pull request #1046 from shelhamer/cudnn

  Show description for 3bafe2f
  shelhamer

  committed
  3bafe2f

  Copy full SHA for 3bafe2f

• Fix more lint.

  jyegerlehner

  committed
  a739cda

  Copy full SHA for a739cda

• Repair crash in conv_layer due to weight pointer being NULL.

  jyegerlehner

  committed
  396da71

  Copy full SHA for 396da71

Commits on Sep 7, 2014

• strategize cuDNN convolution

  shelhamer

  committed
  d1b38ee

  Copy full SHA for d1b38ee

Commits on Aug 13, 2014

• Generalize loss by allowing any top blob to be used as a loss in which

  Show description for 512a626
  jeffdonahue

  committed
  512a626

  Copy full SHA for 512a626

Commits on Aug 7, 2014

• LICENSE governs the whole project so strip file headers

  shelhamer

  committed
  b97b88f

  Copy full SHA for b97b88f

Commits on Aug 5, 2014

• Fix header alphabetization lint errors.

  jeffdonahue

  committed
  92f00dd

  Copy full SHA for 92f00dd

Commits on Jul 29, 2014

• fix GPU indexing

  Show description for ca4482c
  shelhamer

  committed
  ca4482c

  Copy full SHA for ca4482c

• im2col + convolve non-square filters, padding, and stride

  shelhamer

  committed
  4d44fe7

  Copy full SHA for 4d44fe7

Commits on Jul 20, 2014

• use Blob directly instead of shared_ptr for ConvolutionLayer::bias_multiplier_

  Show description for 00ecb24
  longjon

  committed
  00ecb24

  Copy full SHA for 00ecb24

Commits on Jul 17, 2014

• stub out GPU layer methods to crash loudly in CPU-only mode

  shelhamer

  committed
  41e063f

  Copy full SHA for 41e063f

Commits on Jul 15, 2014

• Use Blobs instead of SyncedMemorys for the bias_multiplier_'s.

  jeffdonahue

  committed
  fb8c061

  Copy full SHA for fb8c061

• Make ConvolutionLayer and InnerProductLayer abide by param_propagate_down_

  jeffdonahue

  committed
  d5351bc

  Copy full SHA for d5351bc

Commits on Jul 4, 2014

• replace all memset with caffe_set() / caffe_gpu_set()

  Show description for af624fa
  shelhamer

  committed
  af624fa

  Copy full SHA for af624fa

Commits on Jul 3, 2014

• ConvolutionLayer can take N bottom blobs and N top blobs

  jeffdonahue

  committed
  a2b2874

  Copy full SHA for a2b2874

Commits on Jun 26, 2014

• change Backward interface: propagate_down is a vector -- use to fix

  Show description for 7444359
  

  jeffdonahue

  authored and

  shelhamer

  committed
  7444359

  Copy full SHA for 7444359

Commits on Jun 9, 2014

• layers declare their names and number of input/output blobs, and don't

  Show description for 69af0c8
  jeffdonahue

  committed
  69af0c8

  Copy full SHA for 69af0c8

Commits on Mar 28, 2014

• some cleanup - lowercase layer class member variable names

  jeffdonahue

  committed
  fb25211

  Copy full SHA for fb25211

• changes to layers etc to make 'make all' run successfully under new

  Show description for ecb3873
  jeffdonahue

  committed
  ecb3873

  Copy full SHA for ecb3873

Commits on Mar 27, 2014

• Standardize copyright, add root-level CONTRIBUTORS credit

  shelhamer

  committed
  6db92e8

  Copy full SHA for 6db92e8