lyuwenyu
Results
6
issues of
lyuwenyu
The `downsample` part in each block/layer (not the skip connection part), the PyTorch do it in conv3x3 using stride=2, but official caffe version in conv1x1 with stride=2 ``` conv1x1 ->...
wontfix
module: models
topic: classification
### 解决方式： 1. 复现SQR-Deformable detr-R50的精度，（数据为MS-COCO、模型、代码均已开源）（paddledet有Deformable detr-R50这个模型）。 2. 加入到SQR策略到ppdet，验证精度并说明在提交的PR中。 3. 提交PR并合入到paddledet。 - 论文: [Enhanced Training of Query-Based Object Detection via Selective Query Recollection](https://arxiv.org/abs/2212.07593) - 代码 https://github.com/Fangyi-Chen/SQR - 指标 - 数据
feature request
- move `utils.py`from examples/ to models/