search

hoiliu-0801 / DNTR

A DeNoising FPN with Transformer R-CNN for Tiny Object Detection
Apache License 2.0
19 stars 1 forks source link

The code of Transformer-based models. #6

Closed dyl96 closed 3 weeks ago

dyl96 commented 1 month ago

Hello author, may I ask if it is convenient for you to share the code of Transformer-based models in the paper? For academic purposes and will also cite your paper. Thank you very much！

hoiliu-0801 commented 1 month ago

You may refer to "mmdet-dntr/configs_ori" to see the config of DETR series. The latest mmdet version also supports DINO and Conditional DETR. Our code for the transformer results will be released soon at https://github.com/Katie0723/DQ-DETR. Thank you for the citation.

dyl96 commented 1 month ago

Thank you！May I ask if the Params, FLOPs, and accuracy of different categories of Transformer based methods in the DQ-DETR paper can be reported?

hoiliu-0801 commented 1 month ago

You can run mmdet-dntr/tools/analysis_tools/get_flops.py to calculate them.