-
-
Hi Yiyu Wang,
It looks like the swin_large_patch4_window12_384_22kto1k_no_head.pth file isn't on google drive. Can you upload it there?
Thanks for all your hard work!
-
run `main.py` with args as
```
"args": [
"--backbone", "swin_large",
"--drop-path-rate", "0.5",
"--num-queries-one2one", "900",
…
-
未修改 ’upernet_vitae_win_window7_512x512_80k_potsdam_epoch100.py ‘参数,分别使用
RGB+ Label_all,IRRG + Label_all复现了一遍
我是使用 tools/convert_datasets/potsdam.py 脚本处理的
config, model
结果log和放出的…
-
### Checklist
- [x] I have searched related issues but cannot get the expected help.
- [ ] 2. I have read the [FAQ documentation](https://github.com/open-mmlab/mmdeploy/tree/main/docs/en/faq.md) but …
-
Hi,
For the SWIN encoder in your model, did you use pretrained weights from ImageNet or did you train your network end-to-end from scratch?
Thanks
-
[SSR-Net](https://github.com/shamangary/SSR-Net): A Compact Soft Stagewise Regression Network for Age Estimation
sngyo updated
4 months ago
-
I am tring to reproduce the result for uppernet + swin transformer. However, when I run: python tools/train.py configs/swin/upernet_swin_base_patch4_window12_512x512_160k_ade20k_pretrain_384x384_22K.p…
-
### Checklist
- [X] I have searched related issues but cannot get the expected help.
- [X] 2. I have read the [FAQ documentation](https://github.com/open-mmlab/mmdeploy/tree/main/docs/en/faq.md) b…
-
感谢作者的工作!
我复现NYUv2的训练,Swin-L,4卡P40,每卡batch只能训练1张图,预估训练耗时145h,训练效率正常吗?
目前训练到21个epoch,eval结果NYUv2性能RMSE0.338,而且期间波动还不小到0.38左右。离目标0.314还有不少差距。请问训练epoch影响很大么,以及模型训练稳定性是怎样的?
感谢!