Commits

Branch selector

 master

User selector

All users

Datepicker

All time

Commit History

Commits on Aug 15, 2018

• Merge pull request roytseng-tw#121 from vfdev-5/update_0.4.1

  Show description for 8315af3
  roytseng-tw

  authored
  8315af3

  Copy full SHA for 8315af3

Commits on Aug 6, 2018

• Merge pull request roytseng-tw#118 from ncherel/resume_error

  Show description for 3f533b5
  roytseng-tw

  authored
  3f533b5

  Copy full SHA for 3f533b5

Commits on Aug 4, 2018

• Fix compiling errors for pytorch 0.4.1

  vfdev-5

  committed
  cb237c0

  Copy full SHA for cb237c0

Commits on Jul 31, 2018

• Fix resume error

  

  Nicolas

  committed
  752815d

  Copy full SHA for 752815d

Commits on Jul 24, 2018

• Fix regex pattern for resent weights names in Detectron

  Show description for dddeb6f
  roytseng-tw

  committed
  dddeb6f

  Copy full SHA for dddeb6f

Commits on Jun 10, 2018

• Fix bugs on optimization settings for groupnorm parameters

  roytseng-tw

  committed
  9defe95

  Copy full SHA for 9defe95

Commits on Jun 2, 2018

• Fix issue on comparing state_dict values on pytorch0.4

  roytseng-tw

  committed
  b07775a

  Copy full SHA for b07775a

Commits on May 31, 2018

• Implement augmented testing and fix a bug to ensure no_grad when testing with pytorch0.4

  roytseng-tw

  committed
  0e0ca55

  Copy full SHA for 0e0ca55

Commits on May 27, 2018

• Fix pytorch UserWarning: invalid index of a 0-dim tensor when use '--iter_size n'

  roytseng-tw

  committed
  623e2c4

  Copy full SHA for 623e2c4

Commits on May 25, 2018

• Update README.md about ResNeXt backbone

  roytseng-tw

  committed
  9ab1ea9

  Copy full SHA for 9ab1ea9

• Add all baseline config files of ResNeXt backbone

  roytseng-tw

  committed
  27505d4

  Copy full SHA for 27505d4

• Support ResNeXt backbones and refactor ResNet.py

  roytseng-tw

  committed
  dbba6d4

  Copy full SHA for dbba6d4

Commits on May 24, 2018

• Update issue_template.md

  roytseng-tw

  authored
  23c5c1e

  Copy full SHA for 23c5c1e

Commits on May 22, 2018

• Fix R-101-GN.pkl download link in README.md

  roytseng-tw

  authored
  9259132

  Copy full SHA for 9259132

• Update README.md for group normalization

  roytseng-tw

  authored
  1134c57

  Copy full SHA for 1134c57

• Merge branch 'master' of github.com:roytseng-tw/Detectron.pytorch

  roytseng-tw

  committed
  a4b7e37

  Copy full SHA for a4b7e37

• Update README.md

  Show description for c1f689e
  roytseng-tw

  authored
  c1f689e

  Copy full SHA for c1f689e

Commits on May 21, 2018

• Reorganize config files and add all end-to-end config files

  roytseng-tw

  committed
  5f9346b

  Copy full SHA for 5f9346b

• Add group norm mask-rcnn baselines

  roytseng-tw

  committed
  4705009

  Copy full SHA for 4705009

• Fix group norm layer

  roytseng-tw

  committed
  2a5facd

  Copy full SHA for 2a5facd

Commits on May 20, 2018

• Merge branch 'master' into add_groupnorm

  roytseng-tw

  committed
  32cfbc6

  Copy full SHA for 32cfbc6

• Add group norm layer

  roytseng-tw

  committed
  ef81b70

  Copy full SHA for ef81b70

• Fix bugs on CLS_AGNOSTIC_BBOX_REG (issue#62)

  roytseng-tw

  committed
  1833c71

  Copy full SHA for 1833c71

Commits on May 17, 2018

• Add benchmark results for mask-rcnn and keypoint-rcnn

  roytseng-tw

  committed
  ab028df

  Copy full SHA for ab028df

• Make detectron weight mapping works with extra custom modules

  Show description for 4cef36e
  roytseng-tw

  committed
  4cef36e

  Copy full SHA for 4cef36e

Commits on May 16, 2018

• Add missing group=cfg.RESNETS.NUM_GROUPS argument for subsequent resblocks, which affects ResNeXt architectures.

  roytseng-tw

  committed
  9023e0e

  Copy full SHA for 9023e0e

Commits on May 14, 2018

• Update README about pytorch0.4 support

  roytseng-tw

  committed
  96b22b1

  Copy full SHA for 96b22b1

• Make inference without gradient compatible with PyTorch0.4

  roytseng-tw

  committed
  067bb23

  Copy full SHA for 067bb23

• Solve DeprecationWarning on PyTorch0.4 for parameter init functions

  Show description for 424499a
  roytseng-tw

  committed
  424499a

  Copy full SHA for 424499a

• keep dim when take means on losses and metrics from multigpus

  Show description for a7a1619
  roytseng-tw

  committed
  a7a1619

  Copy full SHA for a7a1619

• Unsqueeze at dim 0 for losses and metrics tensors to ensure proper gathering from multigpu

  roytseng-tw

  committed
  312035d

  Copy full SHA for 312035d

• Explicitly assign a batchsampler that doesn't cast sampler output to int

  roytseng-tw

  committed
  235885b

  Copy full SHA for 235885b

• Add matplotlib in required depencies

  roytseng-tw

  committed
  cd26cc1

  Copy full SHA for cd26cc1

Commits on May 10, 2018

• Update BENCHMARK.md

  roytseng-tw

  authored
  9294ec1

  Copy full SHA for 9294ec1

• Add BENCHMARK.md

  roytseng-tw

  committed
  5f2811f

  Copy full SHA for 5f2811f

Pagination

Previous

Next