irfanICMLL/structure_knowledge_distillation
The official code for the paper 'Structured Knowledge Distillation for Semantic Segmentation'. (CVPR 2019 ORAL) and extension to other tasks.
PythonBSD-2-Clause
Issues
- 9
How can I run build.py in pytorch1.9?
#63 opened by BBBBchan - 7
- 1
- 1
Training from scratch reached the performance similar with that when initializing from ImageNet
#68 opened by Shawn207 - 1
training log
#70 opened by songyang86 - 4
How to run on my own dataset?
#49 opened by eskeron-yxgg - 0
How to apply KD for two input images?
#69 opened by Tejuwi - 1
- 2
- 2
training fails with "RuntimeError: cuda runtime error (11) : invalid argument at THCGeneral.cpp:405"
#43 opened by betogulliver - 0
/data/ckj/SKD-master/libs/_ext/__ext.so: failed to map segment from shared object
#67 opened by asiguan - 0
code error
#66 opened by Liuhao-128 - 0
- 2
Holistic loss
#62 opened by rose-jinyang - 0
FileNotFoundError: [Errno 2] No such file or directory: './ckpt/CS_scenes_39326_0.75.pth'
#64 opened by GeneralJing - 2
Pretrained student?
#58 opened by mbahri - 2
build error
#36 opened by wl082013 - 3
Why the patch size turn out to be half of the original feature map size in CriterionPairWiseforWholeFeatAfterPool?
#51 opened by Ssssseason - 1
- 1
Train script
#55 opened by GeneralJing - 6
Hello, I also cannot reproduce the results
#38 opened by DanceInDark - 0
About the input of discriminator
#59 opened by blessingforever - 5
Pixel-wise loss
#45 opened by duanxuesong - 1
- 4
- 2
about the code
#54 opened by 15757170756 - 1
Puzzled about this loop
#52 opened by NickChang97 - 2
D_loss becomes NAN during training
#28 opened by zhouzg - 2
codes for PAMI version
#33 opened by ArsenLuca - 0
checkpoint
#53 opened by NickChang97 - 2
- 5
TypeError: add_() received an invalid combination of arguments - got (complex, Tensor)
#44 opened by lmunan - 1
- 1
The paper use average pooling to aggregate β × C features in one node to be 1 × C, but the code shows max pooling.
#46 opened by aye0804 - 2
ESPNet training
#32 opened by sungsooo - 3
no models for detection and depth estimation
#40 opened by wl082013 - 2
retrain teacher net
#42 opened by wl082013 - 1
Build error
#37 opened by shidilrzf - 1
CUDA Error encountered in <function CompiledLib.bn_forward_cuda at 0x7fef222d0400>
#41 opened by zhLawliet - 4
Hi, I cannot reproduce the result.
#22 opened by aachenhang - 3
Could you provide your training log?
#35 opened by happycoding1996 - 1
Question about the pair-wise loss
#31 opened by colorjam - 1
The source code of FCOS
#30 opened by miaoshuyu - 3
pixel-wise loss dominates generator training
#34 opened by ArsenLuca - 4
- 2
I can't get your mIoU on cityscapes dataset.
#29 opened by mailhoho - 2
run_train_val.sh: --pi: not found
#26 opened by YBIO - 4
nan loss value
#25 opened by sungsooo - 2
- 1
Where is Discriminator weights
#24 opened by lijiannuist