kuangliu/torchcv

Issues

• For Efficient Graph-Based Image Segmentation, the function remove_small_components does not remove any

  #55 opened 3 years ago by xysue2000
  0

• Key mismatch while loading the model?

  #28 opened 6 years ago by Jumabek
  17

• _Py_ZeroStruct

  #54 opened 4 years ago by Azkarehman
  0

• RuntimeError: expected backend CPU and dtype Float but got backend CPU and dtype Long

  #52 opened 5 years ago by PRAYALANKAR
  3

• Problem on Loss Function [division by zero]

  #35 opened 6 years ago by ahkarami
  4

• SSDLoss with hard negative mining

  #53 opened 5 years ago by sandylaker
  0

• is there paper about FPNSSD?

  #51 opened 5 years ago by liuliu66
  0

• Error when testing on VOC dataset

  #40 opened 6 years ago by bowenc0221
  2

• Corner case of box_nms

  #50 opened 6 years ago by DHPO
  1

• ModuleNotFoundError: No module named 'torchcv'

  #49 opened 6 years ago by hxy1051653358
  1

• COCO evaluation code

  #41 opened 6 years ago by bowenc0221
  0

• Problem in getting anchor box

  #44 opened 6 years ago by AlanStark
  3

• PyTorch Version support

  #47 opened 6 years ago by AlanStark
  0

• why with hard_example_mining use this?

  #46 opened 6 years ago by kweisamx
  0

• Some questions about train

  #45 opened 6 years ago by LiangSiyuan21
  0

• 'FPNSSD512' object has no attribute 'steps'

  #43 opened 6 years ago by nico1as
  0

• Image size while prediciton

  #42 opened 6 years ago by MohitTare
  0

• Is there a manual on how to use and train it?

  #31 opened 6 years ago by yuliiasergeeva
  1

• The code to generate dataset lists

  #26 opened 6 years ago by D-X-Y
  1

• "RuntimeError: seq can't be empty" in demo.py

  #15 opened 6 years ago by nico1as
  2

• Evaluation speed or frame per second？

  #39 opened 6 years ago by zxiaomzxm
  0

• Why not reduce the rgb_mean in the preprocess?

  #38 opened 6 years ago by liuwei16
  0

• Training on new dataset

  #37 opened 6 years ago by ljtruong
  4

• Non Normalized coordinates

  #30 opened 6 years ago by vaishnavm217
  4

• Dimension mismatch in box_coder.py

  #36 opened 6 years ago by chrisliu54
  0

• random_crop results in image with zero target.

  #16 opened 7 years ago by chenyuntc
  2

• Scale Ratios

  #34 opened 6 years ago by ahkarami
  5

• Add Other Detection Models

  #33 opened 6 years ago by ahkarami
  0

• Change Anchor Boxes

  #32 opened 6 years ago by ahkarami
  0

• evaluation

  #14 opened 7 years ago by nationalflag
  8

• Some questions in models.ssd.net.py

  #19 opened 7 years ago by YingdiZhang
  1

• Why do you evaluate on cpu(not on gpu)?

  #29 opened 6 years ago by Jumabek
  1

• device-side assert triggered

  #25 opened 6 years ago by AminAbd
  6

• ssd training cuda problem

  #27 opened 6 years ago by kelly08385
  0

• retina anchor box computation

  #24 opened 6 years ago by to3i
  0

• RuntimeError: dimension out of range (expected to be in range of [-1, 0], but got 1)

  #23 opened 6 years ago by geonseoks
  1

• training strategy？

  #22 opened 7 years ago by YingdiZhang
  0

• Unused Parameters in torchcv/models/ssd/net.py

  #21 opened 7 years ago by rosenfeldamir
  0

• Does your train set contain difficult instances?

  #20 opened 7 years ago by Z-Zheng
  0

• How can I train FPNSSD network?

  #18 opened 7 years ago by wang-tf
  0

• why index < 0 ?

  #17 opened 7 years ago by airyym
  0

• Weights files are missing

  #2 opened 7 years ago by lemairecarl
  3

• Question about run time

  #13 opened 7 years ago by CNSaintNeo
  0

• Question about focal loss impementation

  #12 opened 7 years ago by Tord-Zhang
  0

• About the results on the Leaderboard

  #11 opened 7 years ago by szq0214
  2

• Difference from DSSD?

  #10 opened 7 years ago by Rizhiy
  0

• Why not using models from torchvision.models?

  #6 opened 7 years ago by aosokin
  1

• will release pretrained model?

  #5 opened 7 years ago by jxlijunhao
  1

• Bug: class target can't be negative

  #4 opened 7 years ago by lemairecarl
  2

• Add license

  #3 opened 7 years ago by lemairecarl
  1