Commits

Branch selector

 2b6e320

User selector

All users

Datepicker

All time

Commit History

Commits on May 5, 2015

• 74 make the test work (ony works with float at the moment)

  

  Chris Bamford

  committedMay 5, 2015
  2b6e320

Commits on May 4, 2015

• 74: found huge memory issue, rand function was trying to use host memory.

  

  Chris Bamford

  committedMay 4, 2015
  16400e4

Commits on Apr 30, 2015

• using CublasPointer in lots more places as it wraps the data transfer between device and host much more reliable and is in one single place

  

  Chris Bamford

  committedApr 30, 2015
  0b0fb93

• changing DataBuffer dtype integer to an ENUM so it makes more sense and is less liable to be broken,

  

  Chris Bamford

  committedApr 30, 2015
  f0ca42e

Commits on Apr 29, 2015

• only allocate the buffer that we need, if there is a byte offset in the INDArray then only allocate the data past that point, and only copy past that point also, otherwise it's segfault city

  

  Chris Bamford

  committedApr 29, 2015
  3cc7a00

• sync around cuda buffer

  

  Chris Bamford

  committedApr 29, 2015
  c863ed2

• array out of bounds exception killed

  

  Chris Bamford

  committedApr 29, 2015
  9200d93

• easier to debug test

  

  Chris Bamford

  committedApr 29, 2015
  084e143

• wrong pointer being sent to device

  

  Chris Bamford

  committedApr 29, 2015
  1601cb1

• don't treat column vectors like row vectors if we are accumulating

  

  Chris Bamford

  committedApr 29, 2015
  1e9f9fd

• remove ensure not freed from cuda not a necessary check anymore

  

  Chris Bamford

  committedApr 29, 2015
  f045fb7

• missing copy and frees

  

  Chris Bamford

  committedApr 29, 2015
  55ce4b4

• reverting just a buffer copy as strides are super important here.

  

  Chris Bamford

  committedApr 29, 2015
  15b2b49

• wrap pointers to cublas functions neatly

  

  Chris Bamford

  committedApr 29, 2015
  753b923

• fix assign to use CublasPointer to wrap memory allocation

  

  Chris Bamford

  committedApr 29, 2015
  de76164

• deal with offsets within the kernelparameterswrapper

  

  Chris Bamford

  committedApr 29, 2015
  6302166

• bit of a rename

  

  Chris Bamford

  committedApr 29, 2015
  ac4aa41

• deal with offsets java side and not kernel side

  

  Chris Bamford

  committedApr 29, 2015
  acba76e

• reset the host pointer if we change the buffer

  

  Chris Bamford

  committedApr 29, 2015
  57602a8

• few fixes in the executioner

  

  Chris Bamford

  committedApr 29, 2015
  0dcc0b2

• reduce.h kernels should use the xOffset - PLEASE REVIEW

  

  Chris Bamford

  committedApr 29, 2015
  924b327

• fix dupe for doubles

  

  Chris Bamford

  committedApr 29, 2015
  0c8699f

• distribution tests now passing

  

  Chris Bamford

  committedApr 29, 2015
  37967f1

• fixing uniform distribution test

  

  Chris Bamford

  committedApr 29, 2015
  bd73d8b

• fixing of fast-copy suing underlying databuffers

  

  Chris Bamford

  committedApr 29, 2015
  6071c79

• fixing all the distribution generators with new memory management architecture

  

  Chris Bamford

  committedApr 29, 2015
  4be3967

• move the kernel params tooling out to it's own file to be used elsewhere

  

  Chris Bamford

  committedApr 29, 2015
  100b906

• sort out byte ordering and get rid of orig file

  

  Chris Bamford

  committedApr 29, 2015
  80da8d9

Commits on Apr 28, 2015

• whole load of memory stuff.

  

  Chris Bamford

  committedApr 28, 2015
  8653a31

Commits on Apr 25, 2015

• managing resources manually for cuda, remove buffer reaper or resource manager

  

  Chris Bamford

  committedApr 25, 2015
  02bed42

Commits on Apr 24, 2015

• playing around with using closable resources

  

  Chris Bamford

  committedApr 24, 2015
  beff51b

Commits on Apr 23, 2015

• ensure accumulations dont happen on column vectors with dimension 1

  

  Adam Gibson

  committedApr 23, 2015
  4515246

• Merge branch 'master' of github.com:deeplearning4j/nd4j

  

  Adam Gibson

  committedApr 23, 2015
  7e5dc44

• Merge pull request #83 from tverbele/fix/comparison

  

  Adam Gibson

  committedApr 23, 2015
  d909ddf

• fft a little more sensible output

  

  Adam Gibson

  committedApr 23, 2015
  56fbc83

Pagination

PreviousNext