SuperMHP/GUPNet

Issues

• train训练的时候CPU

  #34 opened a year ago by aqr1961
  5

• 什么时候对depth进行缩放？

  #47 opened 9 months ago by HNzhaoyli
  5

• 关于GUPnet ++和GUPnet的问题

  #45 opened a year ago by benz725
  2

• 关于backbone和neck

  #46 opened 10 months ago by BaronLeeLZP
  1

• vs code 调试时会卡死是什么情况？

  #43 opened a year ago by benz725
  1

• 有两个问题，麻烦解答一下~

  #44 opened a year ago by benz725
  9

• About code for GUPNet++

  #42 opened a year ago by HIT-ShuWei
  1

• i was wonder how to use evaluate code

  #29 opened a year ago by zzy0731
  1

• Inference 在其他dataset 上問題

  #41 opened a year ago by zzy0731
  0

• 关于数据增强的一些疑问

  #40 opened a year ago by wzichen294
  0

• 关于实验可重复性的问题

  #37 opened a year ago by wzichen294
  1

• 关于数据增强的一些疑问

  #39 opened a year ago by wzichen294
  0

• The scale augmentation is not suitable for 3d regression

  #38 opened a year ago by VincentZ-2020
  2

• size3d_loss是负数

  #30 opened a year ago by yjy4231
  1

• 多卡训练报错

  #36 opened a year ago by HNzhaoyli
  1

• pre-trained model for nuscenes

  #35 opened a year ago by stanny880913
  4

• code for GUPNet++

  #33 opened a year ago by yjy4231
  5

• Can you provide the visualization code

  #32 opened a year ago by yjy4231
  0

• Question about train log in validation test

  #28 opened 2 years ago by kaitoud906
  5

• Question about the Calibration.flip/new_calib_matrix-P2

  #31 opened 2 years ago by loipct
  0

• 关于在test上只关注car的结果

  #27 opened 2 years ago by zzHardx
  1

• Much lower AP_3D compared to AP_BEV

  #4 opened 3 years ago by SPengLiang
  11

• Is val mAP=16.46 reported in paper only trained with 'car' category?

  #26 opened 2 years ago by Senwang98
  2

• 关于nn.AdaptiveAvgPool2d，3d网络中加自适应平均池化的作用

  #25 opened 2 years ago by XcodeHw
  1

• Calibration.flip()函数的原理

  #23 opened 2 years ago by kwong292521
  4

• size3d_loss 负数

  #24 opened 2 years ago by YinengXiong
  0

• Zero score on detetion

  #21 opened 3 years ago by pk1996
  1

• How to visualize detected results?

  #19 opened 3 years ago by Grace1109
  1

• Questions about evaluate results

  #20 opened 3 years ago by Senwang98
  3

• urllib.error.HTTPError: HTTP Error 404: Not Found

  #22 opened 3 years ago by chenqiuli0
  1

• About training under different version of pytorch and cuda

  #14 opened 3 years ago by shangbuhuan13
  3

• evaluation scripts?

  #18 opened 3 years ago by dingmiaomiao
  2

• Why does the loss become nan?

  #17 opened 3 years ago by Cc-Hy
  7

• Which script should be used for evaluation in tools? ap11 ? or ap40?

  #8 opened 3 years ago by lji72
  4

• Question About the AP in test set

  #15 opened 3 years ago by yfthu
  1

• question about the uncertainty loss

  #16 opened 3 years ago by tjulyz
  1

• why not the bias depth is a negative number?

  #12 opened 3 years ago by lfydegithub
  3

• Question about the Calibration.flip

  #13 opened 3 years ago by yfthu
  2

• question about affine transformation？

  #9 opened 3 years ago by Senwang98
  1

• why not h3d_std * X + h3d_mean ?

  #11 opened 3 years ago by lfydegithub
  2

• truncated case

  #10 opened 3 years ago by lfydegithub
  5

• Hi, is there a way to get best checkpoint?

  #7 opened 3 years ago by konyul
  1

• Hi, I am not sure about the structure about the KITTI dataset

  #6 opened 3 years ago by RonaldYuren
  1

• Is the network pretrained on a depth set?

  #5 opened 3 years ago by Cc-Hy
  2

• The difference between original paper and released ckpt

  #3 opened 3 years ago by aqianwang
  2

• Question about the release time

  #2 opened 3 years ago by Juzezhang
  1

• Question about the code releasing time

  #1 opened 3 years ago by gujiaqivadin
  1