Commit Graph

  • 989ab8c38a IOU loss function Joseph Redmon 2015-04-24 10:27:50 -0700
  • feabcc31de gonna fuck shit up Joseph Redmon 2015-04-21 12:58:03 -0700
  • 451ef0a0a6 It's time, to du-du-du-du-DU-DU-DUEL!! https://www.youtube.com/watch?v=IVmtUK_1jh4 Joseph Redmon 2015-04-20 08:43:54 -0700
  • f199fd3b64 per image randomness in crop layer Joseph Redmon 2015-04-17 12:32:54 -0700
  • 47528e37cf crop layer scaling and trans on cpu Joseph Redmon 2015-04-15 01:04:38 -0700
  • d97331b88f level adjustment for images Joseph Redmon 2015-04-15 00:32:32 -0700
  • b0106d7bde probably stuff changed Joseph Redmon 2015-04-13 14:09:55 -0700
  • 390a0cf923 not much changed... Joseph Redmon 2015-04-11 01:24:07 -0700
  • 5a49c1d962 rotation Joseph Redmon 2015-04-09 23:00:33 -0700
  • 4f50e29365 big change to images Joseph Redmon 2015-04-09 15:18:54 -0700
  • 1fd10265f8 detection better? Joseph Redmon 2015-04-07 15:25:30 -0700
  • a05c4bd2e9 sqrt Joseph Redmon 2015-04-01 10:25:50 -0700
  • 81751b47dd ..... and back to coords Joseph Redmon 2015-03-30 12:04:03 -0700
  • ad59d7e68e sides of box instead of coords Joseph Redmon 2015-03-30 11:09:01 -0700
  • 6553b3f0e3 no comment Joseph Redmon 2015-03-29 19:31:47 -0700
  • f98bf6bbdb We do our OWN resizing! Joseph Redmon 2015-03-28 16:11:37 -0700
  • cf0300ea55 dropout probably ok Joseph Redmon 2015-03-27 17:32:01 -0700
  • d7d7da2653 Fixed im2col mistake >< face#palm Joseph Redmon 2015-03-26 19:13:59 -0700
  • e92f7d301c smaller gridsize in bias Joseph Redmon 2015-03-24 18:27:12 -0700
  • 56b6561ae4 stuff changed probably Joseph Redmon 2015-03-24 13:20:56 -0700
  • 7100de0b59 going to break stuff Joseph Redmon 2015-03-22 21:28:45 -0700
  • 664c5dd2f2 Subdivisions for batches Joseph Redmon 2015-03-22 09:56:40 -0700
  • 9d418102f4 using caffe's im2col, it's so much better\! Joseph Redmon 2015-03-21 14:17:39 -0700
  • 4af116e996 gonna change im2col Joseph Redmon 2015-03-21 12:25:14 -0700
  • dcb000b553 refactoring and added DARK ZONE Joseph Redmon 2015-03-11 22:20:15 -0700
  • f047cfff99 renamed sigmoid to logistic Joseph Redmon 2015-03-08 11:31:12 -0700
  • 655f636a42 detection layer fixed Joseph Redmon 2015-03-08 11:25:28 -0700
  • 2313a8eb54 Split commands into different files Joseph Redmon 2015-03-06 10:49:03 -0800
  • 26cddc6f93 going to split things up Joseph Redmon 2015-03-06 09:49:05 -0800
  • fb9e0fe336 Big changes to detection Joseph Redmon 2015-03-04 14:56:38 -0800
  • 5f4a5f59b0 captcha stuff Joseph Redmon 2015-02-23 18:52:05 -0800
  • 0f645836f1 Detection is back, baby\! Joseph Redmon 2015-02-10 19:41:03 -0800
  • 979d02126b Generalizing conv layer so deconv is easier Joseph Redmon 2015-02-09 13:27:58 -0800
  • 7ee45082f1 doing some testing w/o rand Joseph Redmon 2015-02-09 13:24:32 -0800
  • 2f62fe33c9 saving weight files as binaries, hell yeah Joseph Redmon 2015-02-06 18:53:53 -0800
  • bfffadc755 Stable place to commit Joseph Redmon 2015-02-04 12:41:20 -0800
  • 0f1a31648c idk, probably something changed Joseph Redmon 2015-01-30 22:05:23 -0800
  • c592fc7491 New name: Darknet Joseph Redmon 2015-01-27 13:52:37 -0800
  • 153705226d Bias updates bug fix Joseph Redmon 2015-01-27 13:31:06 -0800
  • 809f924db2 CUDA so fast Joseph Redmon 2015-01-22 16:38:24 -0800
  • 4ac78c8926 I am so done with opencl, switching to cuda Joseph Redmon 2015-01-20 13:26:46 -0800
  • 6e1d5b45de fast sort of working Joseph Redmon 2015-01-19 22:06:18 -0800
  • 08b757a0bf Stable, needs to be way faster Joseph Redmon 2015-01-14 12:18:57 -0800
  • aa5996d58e Faster Joseph Redmon 2015-01-12 17:27:08 -0800
  • f26da0ad5c Need to fix line reads Joseph Redmon 2014-12-28 09:42:35 -0800
  • 4ab366a805 some fixes, some other experiments Joseph Redmon 2014-12-22 14:35:37 -0800
  • f88baf4a3a 99 problems Joseph Redmon 2014-12-18 15:46:45 -0800
  • 19d3ae7267 message Joseph Redmon 2014-12-18 13:21:30 -0800
  • 47914146d9 working stuffs Joseph Redmon 2014-12-18 11:28:42 -0800
  • ecbeec86bf changed nist testing Joseph Redmon 2014-12-18 11:27:13 -0800
  • 9e98f1b68e Not sure Joseph Redmon 2014-12-18 10:47:33 -0800
  • d6cdea49eb Small fixes Joseph Redmon 2014-12-16 15:41:36 -0800
  • 884045091b lots of cleaning Joseph Redmon 2014-12-16 15:34:10 -0800
  • d6fbe86e7a updates? Joseph Redmon 2014-12-16 11:40:05 -0800
  • 90d354a2a5 fixed dropout >< Joseph Redmon 2014-12-13 12:01:21 -0800
  • 79fffcce3c Better imagenet distributed training Joseph Redmon 2014-12-11 13:15:26 -0800
  • aea3bceeb1 timing code Joseph Redmon 2014-12-08 12:11:04 -0800
  • cb1f33c6ae Fixed race condition in server Joseph Redmon 2014-12-08 11:48:57 -0800
  • d81ab1942b mnist now Joseph Redmon 2014-12-07 23:59:45 -0800
  • a99050f0c8 Some fixes to momentum Joseph Redmon 2014-12-07 20:16:21 -0800
  • 28e2115272 Distributed training Joseph Redmon 2014-12-07 00:41:26 -0800
  • 1edcf73a73 Detection good, split up col images Joseph Redmon 2014-12-03 23:20:29 -0800
  • ff67f03476 Starting on server/client Joseph Redmon 2014-12-03 08:48:07 -0800
  • b77a8f3987 stable Joseph Redmon 2014-11-28 10:38:26 -0800
  • e36182cd8c cleaned up data parsing a lot. probably nothing broken? Joseph Redmon 2014-11-21 15:35:19 -0800
  • 7c120aef23 stable, dropout on gpu Joseph Redmon 2014-11-19 14:03:51 -0800
  • d407bffde9 checkpoint Joseph Redmon 2014-11-18 13:51:04 -0800
  • b13ad6d5fd Can validate on imagenet now Joseph Redmon 2014-11-05 14:49:58 -0800
  • 2b2441313b col2im maybe a little faster Joseph Redmon 2014-10-30 11:28:37 -0700
  • 27d0c922ea im2col actually works now Joseph Redmon 2014-10-29 23:47:50 -0700
  • 1c0fd9bb47 im2col slightly faster Joseph Redmon 2014-10-29 23:26:41 -0700
  • af4e4f92dc getting rid of sub_arrays, nvidia driver memory leak Joseph Redmon 2014-10-27 19:45:06 -0700
  • edbccdfcaf Maybe something changed? Joseph Redmon 2014-10-25 22:04:34 -0700
  • 5c9a773bb6 Maybe some stuff changed Joseph Redmon 2014-10-25 17:35:25 -0700
  • 14303717dc Fast, needs to be faster Joseph Redmon 2014-10-25 11:57:26 -0700
  • 158bb1bee9 softmax on gpu Joseph Redmon 2014-10-21 14:49:18 -0700
  • 9b3c7136f3 Fixing up maxpool layer Joseph Redmon 2014-10-16 15:17:23 -0700
  • 7756cccb79 Refactored connected to use blas Joseph Redmon 2014-10-13 22:31:48 -0700
  • 787d534560 Convolutional working on GPU Joseph Redmon 2014-10-13 00:29:01 -0700
  • 76ee68f96d Trying some stuff w/ dropout Joseph Redmon 2014-08-27 19:11:46 -0700
  • 176d65b765 Nist NIN testing multi-crop Joseph Redmon 2014-08-11 12:52:07 -0700
  • 7add111509 maxpool fixed, good on mnist Joseph Redmon 2014-08-09 08:16:37 -0700
  • d9f1b0b16e probably how maxpool layers should be Joseph Redmon 2014-08-08 12:04:15 -0700
  • b32a287e38 Merge branch 'master' of pjreddie.com:jnet Joseph Redmon 2014-07-17 10:17:52 -0700
  • 076009ebe3 Fixed batch stuff in conv layer Joseph Redmon 2014-07-17 10:14:59 -0700
  • 1b94df24fd Midway through lots of fixes, checkpoint Joseph Redmon 2014-07-17 09:05:07 -0700
  • 70d622ea54 Added batch to col2im, padding option Joseph Redmon 2014-07-13 22:07:51 -0700
  • 0c2cf402af Some small fixes for frame Joseph Redmon 2014-05-09 15:35:58 -0700
  • cd8d53df21 So there WAS this huge bug. Gone now Joseph Redmon 2014-05-09 15:14:52 -0700
  • 5ef74c2031 Slowly refactoring and pushing to GPU Joseph Redmon 2014-05-02 15:20:34 -0700
  • 00d483697a Small updates Joseph Redmon 2014-04-30 16:17:40 -0700
  • 354b0cbdcb GPU flag in Makefile Joseph Redmon 2014-04-19 13:10:59 -0700
  • 324e0a33dd Better alternating between video cards Joseph Redmon 2014-04-17 23:14:13 -0700
  • ac82bde55f New feature extraction method Joseph Redmon 2014-04-17 15:07:00 -0700
  • ec16ee612f Old feature extraction method Joseph Redmon 2014-04-17 15:05:41 -0700
  • b4b729a15e Merge branch 'master' of pjreddie.com:jnet Joseph Redmon 2014-04-17 09:57:30 -0700
  • 5e468d1c13 LRNorm layer, Viz, better cfg Joseph Redmon 2014-04-17 09:51:38 -0700
  • 738cd4c2d7 Visualizations? Joseph Redmon 2014-04-16 17:05:29 -0700
  • cc06817efa Attempt at visualizing ImageNet Features Joseph Redmon 2014-04-11 01:00:27 -0700
  • 2ea63c0e99 Better VOC handling and resizing Joseph Redmon 2014-03-12 21:57:34 -0700