Commits

Breadcrumbs

History for

onmaster

User selector

All users

All time

Commit History

Commits on Jan 8, 2019

• use FLAGS.nb_iters_cls_wmup to control cls. loss's warm-up stage

  jiaxiang-wu

  committed
  7f67fe3

  Copy full SHA for 7f67fe3

• gradually increase the cls. loss's coefficient

  jiaxiang-wu

  committed
  9977260

  Copy full SHA for 9977260

• revise warm-start-related FLAGS variables

  jiaxiang-wu

  committed
  4054fff

  Copy full SHA for 4054fff

• let learner restore weights if *.ckpt files exist

  jiaxiang-wu

  committed
  743e17a

  Copy full SHA for 743e17a

• remove unused FLAGS variables

  jiaxiang-wu

  committed
  06204e9

  Copy full SHA for 06204e9

Commits on Jan 7, 2019

• obtain list of trainable vars in SSD's ModelHelper

  jiaxiang-wu

  committed
  440beeb

  Copy full SHA for 440beeb

• stop passing list of variables to warm_start()

  jiaxiang-wu

  committed
  5aaffaa

  Copy full SHA for 5aaffaa

• only perform non-TF evaluation on the primary worker

  jiaxiang-wu

  committed
  3f2b689

  Copy full SHA for 3f2b689

• evaluate mAP via dump_n_eval()

  jiaxiang-wu

  committed
  60ef68c

  Copy full SHA for 60ef68c

• use <dump_n_eval> for non-TF evaluation operations

  jiaxiang-wu

  committed
  553103f

  Copy full SHA for 553103f

• code style revised

  jiaxiang-wu

  committed
  bd57b58

  Copy full SHA for bd57b58

• remove <params>; use FLAGS instead

  jiaxiang-wu

  committed
  88b7561

  Copy full SHA for 88b7561

• evaluation graph built passed

  jiaxiang-wu

  committed
  53ce63d

  Copy full SHA for 53ce63d

Commits on Jan 6, 2019

• use tf.map_fn to obtain target values & loc. predictions

  jiaxiang-wu

  committed
  c794f96

  Copy full SHA for c794f96

Commits on Jan 4, 2019

• add FLAGS variables for evaluation

  jiaxiang-wu

  committed
  b4af0b9

  Copy full SHA for b4af0b9

• SSD: support specifying the model scope in learners

  jiaxiang-wu

  committed
  5fae99b

  Copy full SHA for 5fae99b

Commits on Jan 3, 2019

• remove unused code; calc cls. accuracy in <metrics>

  jiaxiang-wu

  committed
  9768581

  Copy full SHA for 9768581

• rename tf.summary.scalar variables

  jiaxiang-wu

  committed
  befcc30

  Copy full SHA for befcc30

Commits on Jan 2, 2019

• minor bugs fixed; run passed at local & seven

  jiaxiang-wu

  committed
  c8e5c2c

  Copy full SHA for c8e5c2c

• require list of variables when calling warm_start()

  jiaxiang-wu

  committed
  453b9dd

  Copy full SHA for 453b9dd

• add warm_start() to <AbstractModelHelper>

  jiaxiang-wu

  committed
  16fefaa

  Copy full SHA for 16fefaa

• add ModelHelper for VGG @ Pascal-VOC

  jiaxiang-wu

  committed
  bc7793a

  Copy full SHA for bc7793a

Commits on Dec 29, 2018

• add execution script for VGG @ Pascal-VOC

  jiaxiang-wu

  committed
  387d711

  Copy full SHA for 387d711

Commits on Dec 4, 2018

• Optimized Graph Edit for GPU Inference with *.pb Models (#119)

  Show description for 97614a7
  jiaxiang-wu

  authored
  97614a7

  Copy full SHA for 97614a7

Commits on Nov 22, 2018

• Setup learning rate in each ModelHelper (#90)

  Show description for 8033dd4
  

  jiaxiang-wu

  authored and

  psyyz10

  committed
  8033dd4

  Copy full SHA for 8033dd4

Commits on Oct 31, 2018

• initial commit

  jiaxiang-wu

  committed
  82244d9

  Copy full SHA for 82244d9

Loading