NVlabs SegFormer issues

NVlabs / SegFormer

Official PyTorch implementation of SegFormer

https://arxiv.org/abs/2105.15203

Other

2.36k stars 332 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

can we scale mit to a large model which like swin-large？

#150 swjtulinxi opened 4 days ago
0
Exported custom segformer model with pretrained weights ("nvidia/segformer-b0-finetuned-ade-512-512") to ONNX takes longer for inference compare to PyTorch!

#149 Fazankabir opened 2 weeks ago
0
Implementation issues of the Efficient Self-Attention module

#148 K1t3 opened 2 months ago
0
raise StopIteration

#147 dawang-11 opened 3 months ago
0
how to config mmcv-full-1.2.7 on Windows

#146 honghongym opened 3 months ago
2
[Reward] How to freeze and finetune?

#145 Meoooww opened 5 months ago
0
single-gpu training is ok, but Multi-gpu training has some wrong result

#144 zhedahe opened 5 months ago
0
Update image_demo.py to make demo more easy to use

#143 LWShowTime opened 6 months ago
0
Use of different optimizers in SegFormer

#142 anirbala98 opened 6 months ago
0
Potential mistake in SegFormer model: `patch_size` argument in SegFormer model not being used.

#141 jonasdieker opened 6 months ago
1
RuntimeError: Only tuples, lists and Variables supported as JIT inputs/outputs. Dictionaries and strings are also accepted but their usage is not recommended. But got unsupported type numpy.ndarray

#140 hengyanchen closed 6 months ago
0
Model file not found

#139 daviscsuft opened 7 months ago
0
Could you share pretrain recipe?

#138 sipie800 opened 8 months ago
0
How to save the best checkpoint file?

#137 Meoooww opened 8 months ago
0
Permission Issues with Google Drive Link for MiT ImageNet Weights

#136 lhoyer opened 8 months ago
1
Where can I set parameters such as epoch?

#135 Meoooww opened 9 months ago
1
Questions about the Effective Receptive Field

#134 lcy8417 opened 9 months ago
0
Issue about setting number of GPUs for evaluation during training process

#133 Yonghao-Tan closed 9 months ago
0
k v compression in attention may causes small targets and detail to be lost？

#132 swjtulinxi opened 9 months ago
0
keyError 'data_time'

#131 JJY0710 opened 11 months ago
1
License of SegFormer pretrained weights available on other platforms

#130 alzaia opened 11 months ago
0
ImportError: cannot import name 'print_log' from 'mmcv'

#129 alon2U opened 11 months ago
0
The result of Segformer-b5 on Cityscapes

#128 cht619 opened 1 year ago
0
What is the image size used to calculate the flops in Table 1?

#127 wzhiyan opened 1 year ago
0
SegFormerB1 CityScapes - CUDA error: an illegal memory access was encountered

#126 adriengoleb opened 1 year ago
2
ADE B3 model SS mIoU doesn't match paper Table 1.

#125 gathierry opened 1 year ago
3
TypeError: `model` must be a `LightningModule` or `torch._dynamo.OptimizedModule`, got `SegformerForSemanticSegmentation`ON CLOAB

#124 Wang-taoshuo opened 1 year ago
0
How can Fig. 6 be reproduced?? (Effective Receptive Field)

#123 Loparch opened 1 year ago
1
ModuleNotFoundError: No module named 'mmcv._ext'

#122 bekhzod-olimov opened 1 year ago
2
Welcome update to OpenMMLab 2.0

#121 vansin opened 1 year ago
0
Clarifying the Pretraining Procedure

#120 gauenk closed 1 year ago
1
why does the model have to pre-trained on ImageNet-1k datasets? Did you test results directly on ADE20K without using pre-training?

#119 754467737 opened 1 year ago
1
how to calculate models fps in paper

#118 Shinsungjun opened 1 year ago
1
Incorrect Visualization

#117 Jayden9912 opened 1 year ago
2
Pre-training info

#116 nargenziano opened 1 year ago
5
Adding .contiguous() after transpose or permutation

#115 Jayden9912 opened 1 year ago
0
How to convert the model to tensorrt or onnx?

#114 yuchenlichuck opened 1 year ago
2
Excuse me, how to evaluate the accuracy of Cityscapes dataset？

#113 ZzzTD opened 1 year ago
0
Weights for the SegFormer-B5 on COCOStuff

#112 edwardyehuang closed 1 year ago
0
Dataset Creation

#111 FatemaD1 opened 1 year ago
0
backbone about SegNeXt

#110 kamen007 opened 1 year ago
0
CVE-2007-4559 Patch

#109 TrellixVulnTeam opened 1 year ago
0
Why does the image named val_mIoU so strange?

#108 Buling-Knight opened 1 year ago
1
What does "type = IMTRv21_5" mean?

#107 Buling-Knight opened 1 year ago
3
Problems with Docker on remote development server

#106 mak7ostapenko opened 1 year ago
0
Ade20k dataset val/test problem

#105 gavin-gqzhang opened 1 year ago
1
Geforce RTX 3060 must use torch11.0+(Windows)

#104 Buling-Knight opened 1 year ago
4
Math wrong in Overlap Patch Merging

#103 timdirr opened 1 year ago
0
About the ImageNet1k pretrain

#102 zhanggj821 opened 1 year ago
5
The checkpoint can not be used

#101 25884 opened 1 year ago
0