tztztztztz/eqlv2

Issues

• I train eqlv2 using 8 gpus, howerer, the AP is only 0.192

  #7 opened 4 years ago by CRuJia
  2

• eqlv2.py是不是有语法问题？

  #30 opened a year ago by xiaoche-24
  0

• Why doesn't it work？

  #1 opened 4 years ago by JiangWeiHn
  10

• 你好，我希望在自定义的数据集上使用eqlv2，请问标注格式要如何修改？

  #29 opened 2 years ago by yeyanli
  0

• RuntimeError: The expanded size of the tensor (501) must match the existing size (1204) at non-singleton dimension 1.

  #26 opened 2 years ago by pohunshi
  1

• eqlv2 for softmax multi-classification.

  #27 opened 2 years ago by yan-ctrl
  1

• Welcome update to OpenMMLab 2.0

  #28 opened 2 years ago by vansin
  0

• an error in forward function

  #25 opened 2 years ago by SicongLiu998
  0

• How to train EQL v2 with my imbalance dataset( coco_type format)?

  #24 opened 2 years ago by lyon-v
  3

• installation problem

  #23 opened 2 years ago by twangnh
  0

• EQLv2 + RFS would it work?

  #19 opened 3 years ago by kostas1515
  1

• Implementation of category-agnostic task

  #14 opened 3 years ago by IssacCyj
  3

• about the background categories and Instances and Tasks

  #21 opened 3 years ago by Never-Walk-Away
  2

• EQLv2 for softmax loss would it work?

  #22 opened 3 years ago by pixel-lt
  0

• About use_sigmoid and softmax cross_entropy

  #20 opened 3 years ago by Never-Walk-Away
  2

• how did you draw figure1: gradient ratio？

  #18 opened 3 years ago by sanmulab
  4

• eqlv2 for focal loss

  #9 opened 4 years ago by jiangzhengkai
  2

• Some problems about gradient ratio calculation

  #17 opened 3 years ago by sanmulab
  2

• Implementation on Sigle GPU

  #16 opened 3 years ago by sourabhsugandhi
  0

• IndexError: The shape of the mask [1230] at index 0 does not match the shape of the indexed tensor [1231] at index 0

  #13 opened 3 years ago by sanmulab
  7

• Debug problem

  #15 opened 3 years ago by ahmetkoc11
  1

• Question about the paper

  #12 opened 3 years ago by Phoebe-ovo
  6

• Can you please tell where is the code for EQL2 ? I mean how the idea in your paper is implemented. So many folders in this repo that I get confused.

  #8 opened 3 years ago by ShirleyHe2020
  1

• Reproduce the result of Sigmoid showed in paper

  #11 opened 3 years ago by xinqiaozhao
  1

• RuntimeError: Default process group has not been initialized, please make sure to call init_process_group.

  #10 opened 3 years ago by sanmulab
  3

• About lvis dataset

  #5 opened 4 years ago by U-Help
  1

• eqlv2 for regression? Would it work?

  #6 opened 4 years ago by ttsesm
  2

• Function error

  #4 opened 4 years ago by jefferyZhan
  4

• checkpoint performance problem

  #3 opened 4 years ago by yypurpose
  2

• About baseline model

  #2 opened 4 years ago by U-Help
  4