MichaelFan01/STDC-Seg

Issues

• 无法复现论文中的精度

  #103 opened 9 months ago by Asuna88
  1

• 请教下detailloss中多尺度的问题

  #104 opened 9 months ago by Jasonshao0123
  0

• 论文中的热力图是怎样生成的

  #98 opened 2 years ago by LingsiDS
  3

• miou和maxmiou变成了nan和0

  #54 opened 3 years ago by Dandelion111
  4

• 关于训练达不到精度的问题

  #99 opened 2 years ago by CuZaoo
  6

• 请问有没有细节真值图detail GT 的可视化代码

  #97 opened 2 years ago by alittledou
  2

• 1080Ti下，STDC2-seg75和50，用pytorch测速分别只有35和58，用RT加速后也只有65和133，有没有测速成功的大佬分享下经验

  #96 opened 2 years ago by alittledou
  5

• 关于推理速度的问题

  #101 opened 2 years ago by zzuilcc
  0

• Information

  #102 opened a year ago by EthanAbitbol3
  0

• 关于Cityscapes测试集mIOU为nan的问题

  #89 opened 2 years ago by zi-ming
  2

• Train on Camvid

  #85 opened 3 years ago by odyssey0529
  1

• nice work 但测了下时间有更好的模型

  #67 opened 3 years ago
  3

• 模型保存问题

  #100 opened 2 years ago by WenwenZong
  0

• 关于单机单卡遇上的报错

  #83 opened 3 years ago by yang-stephen
  3

• 关于使用的pytorch版本问题

  #68 opened 3 years ago by XiePeiyao
  4

• Visualization and Test speed

  #64 opened 3 years ago by yangpanquan
  1

• What's the purpose of the multipication of n_classes to label in evaluation code?

  #88 opened 2 years ago by lx-r
  1

• 请教test上的结果

  #72 opened 3 years ago by phk1042791021
  3

• UnicodeDecodeError: 'utf-8' codec can't decode byte 0x87 in position 0: invalid start byte

  #75 opened 3 years ago by GeneralJing
  1

• 想问下大佬们，该代码可以在win10下跑通吗

  #95 opened 2 years ago by guodong110
  0

• 你好，我使用自己的数据集，我使用256*256图片训练，结果对800*800的图片分割效果很好 对小于800的分割效果差

  #57 opened 3 years ago by zwd1993
  4

• 关于图片与视频推理代码

  #76 opened 3 years ago by WangzekunY
  1

• 请问在stdcnet.py文件里面的STDCNet1446最后返回的feat2, feat4, feat8, feat16, feat32, idx0_de, idx1_de, idx2_de后面三个是什么意思？

  #94 opened 2 years ago by john-stars
  0

• imagenet预训练

  #56 opened 3 years ago by mc171819
  1

• How can I train with single GPU ?

  #87 opened 3 years ago by CCodie
  4

• question about pretrained model

  #93 opened 2 years ago by 666jonh666
  0

• Nice work！请问我们减少其中一个类别，在训练过程中的val阶段报错

  #92 opened 2 years ago by hankerbit
  1

• 如何理解trainid=-1

  #86 opened 3 years ago by 774911840
  2

• visulization is not so good

  #81 opened 3 years ago by DRosemei
  3

• No module named 'inplace_abn'

  #69 opened 3 years ago by ZhaoNan1
  10

• Detail 模块中用 dice loss 是否合理

  #91 opened 2 years ago by laocan
  0

• model_stages_trt.py问题

  #90 opened 2 years ago by xinguta001
  0

• TensorRT network must have at least one output and validation failed

  #65 opened 3 years ago by DuyguSerbes
  3

• Output image only person

  #84 opened 3 years ago by myoh97
  0

• How to understand the OhemCEloss?

  #82 opened 3 years ago by feiyuhuahuo
  0

• 请问你们后续是否有支持COCO格式的计划？

  #80 opened 3 years ago by dragonhaha
  0

• Can results without pretraining be published?

  #79 opened 3 years ago by Tracisran
  0

• 关于图片尺寸的问题

  #78 opened 3 years ago by 1huyouquan
  1

• 论文中的Fig 3(a)stage6与Fig 4(a)FFM module疑问

  #77 opened 3 years ago by karta2155802
  0

• STDC backbone for imagenet

  #73 opened 3 years ago by jjxxmiin
  2

• 使用STDC1seg预训练权重训练问题

  #74 opened 3 years ago by Mobu59
  0

• ImportError: No module named 'inplace_abn'

  #63 opened 3 years ago by chen1234520
  12

• Issues about loading pretrained model.

  #71 opened 3 years ago by Windaway
  0

• 训练问题

  #70 opened 3 years ago by XMZXW
  0

• 关于STDC模块的疑问

  #66 opened 3 years ago by whuhit
  3

• Reconstruction of test image

  #62 opened 3 years ago by DuyguSerbes
  2

• Can you provide a script to do prediction on single image with the trained model?

  #60 opened 3 years ago by weiningwei
  1

• Cannot run evaluation.py

  #58 opened 3 years ago by drCJH
  3

• 关于Detail Ground-truth Generationd 的问题

  #59 opened 3 years ago by 1huyouquan
  0

• 关于loss计算

  #55 opened 3 years ago by whyandbecause
  0