PeizeSun SparseR-CNN issues

PeizeSun / SparseR-CNN

[CVPR2021, PAMI2023] End-to-End Object Detection with Learnable Proposal

MIT License

1.32k stars 187 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

I am just start to learn, I found that the default configuration of Sparse R-CNN had better AP results than Faster R-CNN. However, with the detection results, I noticed that Sparse R-CNN had lower confidence scores for its detections. Why is that the case? im little confused

#113 ZhangShutongo opened 6 months ago
0
请问怎么使用预训练权重，从头开始训练的准确率比较低

#112 WinnieMs opened 8 months ago
0
How to visualize the activation maps of dynamic head in Sparse R-CNN？

#111 Nanakl opened 1 year ago
0
Initialization box selection

#110 YH-2023 opened 1 year ago
0
Some questions about positional encoding, in the Table 10

#109 AlterRay opened 1 year ago
0
The results were not reproduced on CrowdHuman dataset

#108 a1b2c3s4d4 opened 1 year ago
1
"Dataset 'CrowdHuman_train' is not registered

#107 Aliweka2020 opened 2 years ago
3
KeyError: 1, when using only 1 class dataset to evaluation

#106 YipChakwai opened 2 years ago
0
About performance of Sparse R-CNN

#105 ty1234567454 closed 2 years ago
1
关于动态头的级联问题

#104 Xavier-Lin closed 2 years ago
0
refactoring code with var unpack

#103 idiomaticrefactoring opened 2 years ago
1
AssertionError: Attribute 'thing_classes' in the metadata of 'coco_2017_train' cannot be set to a different value!

#102 Aruen24 opened 2 years ago
2
检测不了物体

#101 ztyf-lq closed 2 years ago
1
Why the optimization targets of refined bboxes is assigned by refined box instead of unrefined bbox?

#100 Mingfeng-Wang closed 2 years ago
2
How can I use a custom dataset?

#99 JEO-96 opened 2 years ago
2
我把从coco数据集换成了自己的数据集，但是出现了错误，我应该修改你的代码中的哪几个部分的代码

#98 libingDY opened 3 years ago
5
How to install the Sparse R-CNN(it seems amusing)?

#97 xiaomin0416 opened 3 years ago
1
Confusing about the performance in the paper

#96 YJHMITWEB closed 3 years ago
3
How to do use test time augmentation?

#95 bravezzzzzz opened 3 years ago
1
Change ADAMW to SGD

#94 mxin262 closed 3 years ago
4
Does anything changes in detectron2？

#93 kuiran opened 3 years ago
1
Question about the proposal_boxes in the code

#92 botaoye closed 3 years ago
2
Resume started from zero again!

#91 zobeirraisi opened 3 years ago
3
How the feature re-use is performed

#90 XuyangBai opened 3 years ago
1
'assert (boxes1[:, 2:] >= boxes1[:, :2]).all()' happened when training

#89 Swall0w opened 3 years ago
4
how to visualize train log?

#88 Yixiang-Wu closed 3 years ago
0
how to use plot function?

#87 Yixiang-Wu opened 3 years ago
1
Question about the evaluation

#86 FL77N closed 3 years ago
10
Init proposal box Visualization

#85 junha1125 closed 3 years ago
2
Question about the Iterative architecture

#84 GuangxingHan opened 3 years ago
2
About the loss format

#83 FL77N opened 3 years ago
4
train w/o pre-trained backbone

#82 StarsMyDestination opened 3 years ago
3
question about pred_bboxes.detach() in DynamicHead.forward()

#81 burningisnaive opened 3 years ago
2
the learning rate adjustment when training with one gpu or two gpus

#80 Young0111 opened 3 years ago
5
保存模型与提供模型大小之间的差异

#79 luyao-cv closed 3 years ago
1
Multi-head attention in dynamic head

#78 HYUNJS closed 3 years ago
3
Question regarding proposal feature

#77 abeyang00 opened 3 years ago
4
Create __init__.py

#76 dashu233 closed 3 years ago
0
Hi,Have you verified the model to ONNX, i meet some bugs to onnx.

#75 csliuchang opened 3 years ago
2
Does input data annotations need to be normalized coordinate before training?

#74 sonnguyenasu closed 3 years ago
2
Will you release DetCo code and model version?

#73 Mingfeng-Wang closed 3 years ago
3
issues about Mix-Precision training

#72 changewOw opened 3 years ago
1
how to do Grid and Rondom proposal boxes initialization?

#71 liz6688 opened 3 years ago
5
Detco

#70 iFighting closed 3 years ago
0
detco pretrain support

#69 iFighting closed 3 years ago
0
When set num_heads = 1, doesnt train

#68 abeyang00 opened 3 years ago
12
cuda out of memory issue

#67 abeyang00 closed 3 years ago
0
Have you tried smooth l1 + giou?

#66 Mingfeng-Wang closed 3 years ago
1
question about learned proposal boxes

#65 liyunsheng13 opened 3 years ago
1
在小数据集上for example （train:60 test:13）如何调优参数，使用默认配置测试结果不太好

#64 csliuchang closed 3 years ago
1