lxtGH/SFSegNets

Issues

• 使用训练好的模型也需要mapillary数据吗

  #80 opened 3 months ago by kll-df
  0

• 请问怎样测试速度

  #79 opened 6 months ago by takaniwa
  1

• SFNet-Lite DSN issue [Not using dsn in forward pass]

  #78 opened 5 months ago by jnisarg
  1

• Does anybody successfully convert this model to TensorRT?

  #72 opened 6 months ago by dlz620301
  5

• SenseTime's reproduction

  #75 opened 6 months ago by sudeqian
  1

• 您好，请问为什么SRNet的模块ModuleHead, ChannelReasonModule的代码不在呢？我很想应用一下，请问您可以释放一下这部分的代码吗？

  #77 opened 6 months ago by takaniwa
  1

• Trying to resize storage that is not resizable

  #76 opened 10 months ago by ygyleo
  0

• FileNotFoundError: [Errno 2] No such file or directory: './data/mapillary/config.json'

  #74 opened a year ago by jingyanliao
  1

• 请问训练的时候是加载单张图片进入模型吗？

  #73 opened a year ago by qubewx
  2

• “对齐”问题求教

  #71 opened 2 years ago by ChengSongyang
  1

• 请问怎么理解“对齐”

  #70 opened 2 years ago by ChengSongyang
  2

• SFNet-Lite: Gated Dual Flow Alignment Module

  #69 opened 2 years ago by yongjin120
  3

• How can I find the code about"Towards Efficient Scene Understanding via Squeeze Reasoning"

  #68 opened 2 years ago by Yiir-1
  2

• No module named "UperNetAlignHeadAddition" in file "network/sfnet_stdcnet.py"

  #66 opened 2 years ago by JianmingLi-888
  1

• 请问为啥sfnet-lite代码和论文结构图不一致？

  #67 opened 2 years ago by swjtulinxi
  4

• Why did you remove the Resnet-101 Cityscape weights ?

  #63 opened 2 years ago by GaetanBaert
  2

• Gridsampler to TensorRT

  #64 opened 2 years ago by jyang68sh
  3

• nan for mIOU

  #56 opened 2 years ago by niloofarAzari
  3

• GPU not being used in eval Cityscapes?

  #65 opened 2 years ago by cuizy3
  1

• Thanks for your great work and I need some help about how to resume training from the latest checkpoint?

  #62 opened 2 years ago by alanlyawen
  0

• Train on a single GPU

  #60 opened 2 years ago by nur-917Hub
  1

• --arch selected_architecture

  #61 opened 2 years ago by m828
  1

• Train on less than 8 GPUs

  #59 opened 2 years ago by BabaVegato
  1

• mapillary-vistas-dataset_public_v1.1.zip

  #58 opened 2 years ago by wmnj
  1

• about config.json

  #55 opened 2 years ago by zhouzhennan
  1

• pretrain for resnet18

  #57 opened 2 years ago by jyang68sh
  1

• different classes lie in different groups along channels sparsely

  #54 opened 3 years ago by whuhit
  3

• Error: CUDA out of memory

  #38 opened 3 years ago by daixiaolei623
  16

• there is not DeepR18_SF_deeply_map network

  #43 opened 3 years ago by daixiaolei623
  1

• C盘的占用率是100%

  #52 opened 3 years ago by gaowq2017
  1

• The "FAM" is indeed a stripped-down version of "DCN"

  #53 opened 3 years ago by heziliustc
  2

• Test mIoUs are better than val mIoUs

  #51 opened 3 years ago by ShihuaHuang95
  1

• transform:failed to synchronize: cudaErrorAssert: device-side assert triggered

  #41 opened 3 years ago by Wu-ZJ
  2

• Papers with Code Leaderboard

  #50 opened 3 years ago by RolandGao
  1

• should i use multigrid[4, 8 16] in SRNet ? the paper use it ,but the code don't use?

  #49 opened 3 years ago by swjtulinxi
  1

• 请问设个参数class_uniform_pct是什么意思？

  #48 opened 3 years ago by swjtulinxi
  1

• Question about the flow_make and flow_warp functions.

  #47 opened 3 years ago by wuwenbin970731
  1

• Dilated ResNet in SFNet?

  #46 opened 3 years ago by wondervictor
  4

• License?

  #44 opened 3 years ago by Jason93K
  2

• question about fps

  #40 opened 3 years ago by songqi-github
  1

• export sfnet_r18_78.pth to onnx

  #45 opened 3 years ago by fjremnav
  1

• Demo still invalid, Please provide pretrained weigth

  #33 opened 3 years ago by lucasjinreal
  6

• also question about fps

  #42 opened 3 years ago by daixiaolei623
  2

• FAM模块在反向传播的时候权重会更新吗

  #39 opened 3 years ago by dd1ge
  2

• Unable to reproduce train

  #37 opened 3 years ago by lucasjinreal
  6

• Unable to train

  #36 opened 3 years ago by lucasjinreal
  14

• How did u train sfnet_resnet_dsn??

  #35 opened 3 years ago by lucasjinreal
  2

• about align_corners in grid_sample

  #34 opened 3 years ago by lingorX
  1

• 直接打印模型结构和提供的训练好的参数对应不太一致？

  #32 opened 3 years ago by hanjr92
  4

• Pooling in inference mode

  #31 opened 3 years ago by rd-wei
  1