Code

Fireflame#1

Code

Merged

jiasenlu merged 15 commits intomasterjwyang/faster-rcnn.pytorch:masterfrom

fireflamejwyang/faster-rcnn.pytorch:fireflameCopy head branch name to clipboard

Aug 30, 2017

Commits

Commits on Aug 25, 2017

• use pretrained caffe model

  jwyang

  committed
  0a7220e

  Copy full SHA for 0a7220e

• add all file

  jwyang

  committed
  b185159

  Copy full SHA for b185159

Commits on Aug 27, 2017

• initial push

  jwyang

  committed
  71a28b1

  Copy full SHA for 71a28b1

• add ratio batch loader

  jiasenlu

  committed
  dce4bf3

  Copy full SHA for dce4bf3

Commits on Aug 29, 2017

• batching data according to aspect ratios, and training in batch

  jwyang

  committed
  b0c7219

  Copy full SHA for b0c7219

• batching data according to aspect ratios, and training in batch

  jwyang

  committed
  20121d4

  Copy full SHA for 20121d4

• minor change to proposal_layer

  jwyang

  committed
  71979d5

  Copy full SHA for 71979d5

• fix a bug in anchor_target_layer, make it work to train in batch

  jwyang

  committed
  0e7da2d

  Copy full SHA for 0e7da2d

• Delete roibatchLoader_aug.py

  jiasenlu

  authored
  46b2d1e

  Copy full SHA for 46b2d1e

Commits on Aug 30, 2017

• change faster_rcnn arch to support multiple base network

  jiasenlu

  committed
  f90836d

  Copy full SHA for f90836d

• Merge branch 'fireflame' of https://github.com/jwyang/graphdetection into fireflame

  jiasenlu

  committed
  913b2f5

  Copy full SHA for 913b2f5

• add vg and imagenet, change trainval

  jwyang

  committed
  a32f872

  Copy full SHA for a32f872

• fix rand bug that cause GPU error. It seems a torch lib bug. use numpy instead

  jiasenlu

  committed
  6def460

  Copy full SHA for 6def460

• Merge branch 'fireflame' of https://github.com/jwyang/graphdetection into fireflame

  jiasenlu

  committed
  9383cf5

  Copy full SHA for 9383cf5

• minor change to vg, make it compatible to roidb

  jwyang

  committed
  58b7d01

  Copy full SHA for 58b7d01