Commits

Branch selector

 master

User selector

All users

Datepicker

All time

Commit History

Commits on Oct 23, 2019

• Merge pull request #49 from torrvision/dependabot/pip/pillow-6.2.0

  Show description for 34f1bd4
  bayesianbrad

  authored
  34f1bd4

  Copy full SHA for 34f1bd4

• Bump pillow from 4.1.1 to 6.2.0

  Show description for 58fec50
  dependabot[bot]

  authored
  58fec50

  Copy full SHA for 58fec50

Commits on Jun 17, 2017

• Merge pull request #3 from PeterHuang2015/patch-1

  Show description for 59cbf83
  bertinetto

  authored
  59cbf83

  Copy full SHA for 59cbf83

• Update README.md

  Show description for 2a6c314
  

  Peter Huang

  authored
  2a6c314

  Copy full SHA for 2a6c314

Commits on Jun 14, 2017

• fixed link

  

  lbertinetto

  committed
  0e6d6d5

  Copy full SHA for 0e6d6d5

Commits on Jun 12, 2017

• Added Precisions AUC performance measure, which reports the integral of center precision error at 50 thresholds between 0 and 25 pixels

  

  lbertinetto

  committed
  a26744c

  Copy full SHA for a26744c

Commits on Jun 10, 2017

• minor

  

  lbertinetto

  committed
  46e985b

  Copy full SHA for 46e985b

• drafted README

  

  lbertinetto

  committed
  0b8465c

  Copy full SHA for 0b8465c

• minor

  

  lbertinetto

  committed
  5dd13d2

  Copy full SHA for 5dd13d2

• rolled-back to original bbox indexing as in groundtruth

  

  lbertinetto

  committed
  ef1bb43

  Copy full SHA for ef1bb43

Commits on Jun 8, 2017

• -1 to all <x,y> from grountruth.txt

  

  lbertinetto

  committed
  f028cb6

  Copy full SHA for f028cb6

• added file to keep track of performance with commits

  

  lbertinetto

  committed
  858aece

  Copy full SHA for 858aece

Commits on Jun 7, 2017

• new parameter for evaluation

  

  lbertinetto

  committed
  fd91a8f

  Copy full SHA for fd91a8f

• -1 to gt also for TC sequences

  

  lbertinetto

  committed
  a6e3960

  Copy full SHA for a6e3960

• added OTB-style precision metric (20 pixels as default)

  

  lbertinetto

  committed
  fce7f30

  Copy full SHA for fce7f30

• More PEP8 compliant

  

  lbertinetto

  committed
  126062e

  Copy full SHA for 126062e

• remove comments

  

  lbertinetto

  committed
  6eb79d2

  Copy full SHA for 6eb79d2

• allow empty custom params

  

  lbertinetto

  committed
  7e828d1

  Copy full SHA for 7e828d1

• removed unused dir

  

  lbertinetto

  committed
  7938470

  Copy full SHA for 7938470

Commits on Jun 1, 2017

• Improve IOU

  Show description for 691754a
  jvlmdr

  committed
  691754a

  Copy full SHA for 691754a

Commits on May 30, 2017

• added requirements

  bertinetto

  committed
  eef5d7e

  Copy full SHA for eef5d7e

• casting images to tf.float32 goes 39.6 fps on luca's machine

  bertinetto

  committed
  e93c33b

  Copy full SHA for e93c33b

• removed notebooks from repo

  bertinetto

  committed
  a47f725

  Copy full SHA for a47f725

• Merge branch 'master' into single_z

  bertinetto

  committed
  6e1df9c

  Copy full SHA for 6e1df9c

Commits on May 29, 2017

• Added TRE evaluation

  bertinetto

  committed
  812e726

  Copy full SHA for 812e726

Commits on May 25, 2017

• 42.5 by doing score_ = score_/sum(score_

  bertinetto

  committed
  d5cc142

  Copy full SHA for d5cc142

• finished evaluation, but IOU is only 41.0, vs 48.5 of Matconvnet

  bertinetto

  committed
  f6e6dc2

  Copy full SHA for f6e6dc2

Commits on May 23, 2017

• Getting frame_sz with tf.shape

  bertinetto

  committed
  be52a25

  Copy full SHA for be52a25

• removing the padding with average doubles the speed!

  bertinetto

  committed
  1aae8fe

  Copy full SHA for 1aae8fe

Commits on May 22, 2017

• working on run evaluation, graph should be defined only once outside everything

  bertinetto

  committed
  1e665dc

  Copy full SHA for 1e665dc

• minor

  bertinetto

  committed
  d5daffd

  Copy full SHA for d5daffd

• solved problem of doing sess.run twice per iteration, still slow

  bertinetto

  committed
  e06b85d

  Copy full SHA for e06b85d

• bug with filereader because you are running two sessions per frame

  bertinetto

  committed
  a479bc6

  Copy full SHA for a479bc6

Commits on May 20, 2017

• for some reason now is only doing 8fps instead of 20-25

  bertinetto

  committed
  df13e49

  Copy full SHA for df13e49

• computing a single z and then concatenating activations is much slower than computing the activations for 3 times the same crop

  bertinetto

  committed
  f73b996

  Copy full SHA for f73b996

Pagination

Previous

Next