Commits

Branch selector

 master

User selector

All users

Datepicker

All time

Commit History

Commits on Jul 25, 2016

• readme for eccv

  zhirongw

  committedJul 25, 2016
  33a08a7

Commits on Mar 8, 2016

• data whitening for rgb input

  zhirongw

  committedMar 8, 2016
  66a94f8

Commits on Mar 3, 2016

• rgb --> yuv

  zhirongw

  committedMar 3, 2016
  a04254a

• mvn loss fixed

  zhirongw

  committedMar 3, 2016
  ea5a19d

Commits on Mar 1, 2016

• defaulting training parameters changed

  zhirongw

  committedMar 1, 2016
  8d6bb39

Commits on Feb 28, 2016

• fix gmm stability

  zhirongw

  committedFeb 28, 2016
  1bd120d

Commits on Feb 21, 2016

• small fix

  zhirongw

  committedFeb 21, 2016
  7de19e8

Commits on Feb 15, 2016

• various fix for master

  zhirongw

  committedFeb 15, 2016
  ec098de

Commits on Feb 4, 2016

• adding input noise

  zhirongw

  committedFeb 4, 2016
  34fe570

Commits on Feb 2, 2016

• border issues

  zhirongw

  committedFeb 2, 2016
  e26f295

Commits on Jan 26, 2016

• sampling code for 4 neighbors

  zhirongw

  committedJan 26, 2016
  affb67d

Commits on Jan 25, 2016

• 4d lstm sweeps the image by two runs

  zhirongw

  committedJan 25, 2016
  2a73ccc

• add spatial weighted loss option

  zhirongw

  committedJan 25, 2016
  bd9100f

Commits on Jan 24, 2016

• fix bugs for sampling in 3d

  zhirongw

  committedJan 24, 2016
  12d762a

Commits on Jan 23, 2016

• Merge branch 'master' of https://github.com/zhirongw/pixel-rnn

  zhirongw

  committedJan 23, 2016
  080091d

• reorganize sampling code.

  zhirongw

  committedJan 23, 2016
  30b5da9

• reorganize sampling code.

  zhirongw

  committedJan 23, 2016
  6e9e3c9

Commits on Jan 22, 2016

• optimize speed for 3 channel mixture of gaussians

  zhirongw

  committedJan 22, 2016
  123dd85

Commits on Jan 21, 2016

• 3 neighbor pixels model

  zhirongw

  committedJan 21, 2016
  7fcf019

Commits on Jan 19, 2016

• Add LSTMs for 3, and 4 neighbors, and various training utilities

  zhirongw

  committedJan 19, 2016
  9e5810e

Commits on Jan 18, 2016

• fixed a bug for 2d-lstm, and shifts to new scheme of defining neighbors

  zhirongw

  committedJan 18, 2016
  972964b

• bugs for sample code fixed. Still need to test it for the new scheme.

  zhirongw

  committedJan 18, 2016
  ba576b1

Commits on Jan 16, 2016

• optimize speed for mixture of gaussians

  zhirongw

  committedJan 16, 2016
  b3caa53

Commits on Jan 15, 2016

• patch -> patch loss, instead of the last pixel

  zhirongw

  committedJan 15, 2016
  d5cef66

• image sampling

  zhirongw

  committedJan 15, 2016
  5303875

Commits on Jan 13, 2016

• add a sample texture

  zhirongw

  committedJan 13, 2016
  c9915d0

• implement my own mvn

  zhirongw

  committedJan 13, 2016
  34947eb

• DataLoader for a single image, training on single texture

  zhirongw

  committedJan 13, 2016
  158a185

Commits on Jan 12, 2016

• change the output of a sequence to be a single pixel rather than a whole patch

  zhirongw

  committedJan 12, 2016
  ff44fc5

• delete the field for borders

  zhirongw

  committedJan 12, 2016
  eac29f4

• merge gmm decoding and loss to a single loss layer

  zhirongw

  committedJan 12, 2016
  1d21ea5

Commits on Jan 11, 2016

• fix for the grayscale image

  zhirongw

  committedJan 11, 2016
  44dce67

• 2d lstm + GMMs

  zhirongw

  committedJan 11, 2016
  1bc4551

Commits on Dec 7, 2015

• Initial commit

  zhirongw

  committedDec 7, 2015
  c71c9e8