Are you sure you want to delete this task? Once this task is deleted, it cannot be recovered.
sguokaiwei@163.com 06ed213c88 | 1 year ago | |
---|---|---|
.. | ||
configs | 1 year ago | |
datasets | 1 year ago | |
models | 1 year ago | |
tools | 1 year ago | |
util | 1 year ago | |
README.md | 1 year ago | |
engine.py | 1 year ago | |
main.py | 1 year ago | |
requirements.txt | 1 year ago |
The PyTorch implementation of the paper D^2ETR: Decoder-Only DETR with Computationally Efficient Cross-Scale Attention
Model | Epoch | GFLOPs | Params | AP |
---|---|---|---|---|
D^2ETR | 50 | 82 | 35 | 43.2 |
Deformable D^2ETR | 50 | 93 | 40 | 50.0 |
Reuirements and Instation
pip install -r requirements.txt
Training
GPUS_PER_NODE=8 ./tools/run_dist_launch.sh 8 ./configs/ddetr.sh \
--coco_path /path/to/coco \
--pvt_resume /path/to/pvt
GPUS_PER_NODE=8 ./tools/run_dist_launch.sh 8 ./configs/def_ddetr.sh \
--coco_path /path/to/coco \
--pvt_resume /path/to/pvt
Evaluation
GPUS_PER_NODE=8 ./tools/run_dist_launch.sh 8 ./configs/ddetr.sh \
--coco_path /path/to/coco \
--resume /path/to/model \
--eval
GPUS_PER_NODE=8 ./tools/run_dist_launch.sh 8 ./configs/def_ddetr.sh \
--coco_path /path/to/coco \
--resume /path/to/model \
--eval
@misc{lin2022d2etr,
title={D^2ETR: Decoder-Only DETR with Computationally Efficient Cross-Scale Attention},
author={Junyu Lin and Xiaofeng Mao and Yuefeng Chen and Lei Xu and Yuan He and Hui Xue},
year={2022},
eprint={2203.00860},
archivePrefix={arXiv},
primaryClass={cs.CV}
}
No Description
Python Text Cuda C++ Shell other
Dear OpenI User
Thank you for your continuous support to the Openl Qizhi Community AI Collaboration Platform. In order to protect your usage rights and ensure network security, we updated the Openl Qizhi Community AI Collaboration Platform Usage Agreement in January 2024. The updated agreement specifies that users are prohibited from using intranet penetration tools. After you click "Agree and continue", you can continue to use our services. Thank you for your cooperation and understanding.
For more agreement content, please refer to the《Openl Qizhi Community AI Collaboration Platform Usage Agreement》