SSOD

SSOD copied to clipboard

Couldn‘t reproduce the results in the paper

1

Following readme construction, I only got AP 23.5 after training on Seed 1 Percent 1 Coco standard while 25.81 ± 0.28 is claimed in the paper. Here's my setting: Dataset:...

Dopamine0717

Cannot reproduce result for coco

4

Following readme construction, I only got AP 0.065 after training on Seed 1 Percent 1 Coco standard while 25.81 ± 0.28 (+16.76) is claimed in the paper. Here's my setting:...

KaiWU5

AttributeError: 'SemiEvalHook' object has no attribute 'evaluation_flag'

3

Ss-shuang123

error during training

1

Traceback (most recent call last): File "examples/train/train.py", line 193, in main() File "examples/train/train.py", line 182, in main train_detector( File "/app/jomedhi/SSOD-main/mmdet_extension/apis/train.py", line 164, in train_detector runner.run(data_loaders, cfg.workflow) File "/usr/local/lib/python3.8/dist-packages/mmcv/runner/epoch_based_runner.py", line 136,...

joydeepmedhi

关于参数设置对结果的影响？

1

 1、表13给出了不同数据集使用了不同的参数设置，如果想在自己的数据集上调优，请问优先调整哪些参数呢？ 2、训练半监督之前需要先进行有监督模型的训练，请问为什么半监督用了8卡，而有监督只用了2卡训练，如果有监督用8卡是否预训练模型指标更高呢？还是说LabelMatch方法需要的有监督预训练模型不需要很强时，性能提升比使用更强的预训练模型时，指标更高？ 最后感谢你们开源的优秀工作！

moontsar

[email protected]

1

I got FileNotFoundError: [Errno 2] No such file or directory: './dataset/coco/annotations/semi_supervised/[email protected]' Error, thus I would like to ask if I can get this file. thank you!

Ryoo72

labelmatch_hooks eval_score_thr函数报错

4

https://github.com/hikvision-research/SSOD/blob/f914ff28734dc27ff8fba00a1792f2271a85679a/mmdet_extension/core/hooks/labelmatch_hooks.py#L152 当某一类型的预测结果为空时，此部分会报错，list out of index。

buaachen1993

找来找去没找到分布损失的代码，是我看漏了吗？请各位指教

ck6698000

Thanks for your contribution of the SSOD, do you have the plan to release the source only model of the UDA settings? I think they can be treated as the...

ygzhu

how to merge new version mmdet

1

Thanks for your amazing project. I'm trying to apply DeformDETR in your framework. But the DeformDETR was first implemented in mmdet V2.12.0, which is a newer version than yours. Could...

1224wxwx