-
### Question
Hi @xiuqhou,
Niels here from the open-source team at Hugging Face. I discovered your work through a paper that builds on your framework: https://huggingface.co/papers/2407.11699. I wo…
-
### Question
Thanks again for your research! The position relation idea is very smart. I've had success improving DINO/ DDQ / Align DETR baselines with Relation DETR for images with 1500 objects, e.g…
-
### Question
Hi, Thanks for the awesome repo. I am trying to finetune your model on a custom dataset. My GPU memory is not enough to finetune relation_detr_focalnet_large_lrf_fl4_1200_2000.py version…
-
### Bug
Traceback (most recent call last):
File "D:\BaiduNetdiskDownload\Relation-DETR-main\main.py", line 213, in
train()
File "D:\BaiduNetdiskDownload\Relation-DETR-main\main.py", line …
-
### Question
Great work! Your research is very impressive and I've been using it for tiny object detection with great results -- there are many fewer false positives.
However, I'm struggling with …
-
### Question
I noticed that you use batch_size=10 in your paper, while the common setting of batch size is 16. So did you explore this point? By the way, as mentioned in your paper, H-DETR does har…
-
### Question
Hi @xiuqhou ,I wanted to train Relation_detr on my custom coco_dataset but getting error.
![image](https://github.com/user-attachments/assets/48bd084b-53a6-4cc2-9ead-cff33c8a929e)
how …
-
### Question
您好,感谢您出色的工作!Relation DETR和Co DETR都是可以在DINO的基础上通过加入即插即用的模块提升较多性能,COCO 1x性能分别能达到51.7和52.1,请问你们有尝试过在Co DETR的框架基础上加入你们提出的模块吗,性能是否有进一步提升呢?
另一个问题是论文中仅提供了r50和swinL直接在COCO上训练的性能,Co-DETR使用VitL主干…
-
### Question
您好,我想基于relation detr做一些视频的任务,对你的框架上做了一些更改,数据加载更换成了cocoVID,帧作为bs,target只读1帧,目前模型方面使用encoder对多帧建模,decoder还是单帧,训练loss下降正常,测试都正常,可是ap全为0,可能某些关键的接口可能不太适配?请问有哪些需要注意的地方吗,非常感谢!!
### 补充信息
_No…
-
Using the [checkpoint](https://drive.google.com/file/d/1vEN-Nat8CtrqP9LoKj0llgUJjGoI7CX7/view) for visual genome, I get the following errors when I try run `evaluate_egtr.py`. The error relates to mis…