-
mmdetection v1.0 cascade rcnn模型 转成v2后finetune，与直接在mmdet v1.0 finetune对比，发现v2版本 loss刚开始为8.1，第一个epoch后map 为74.8。
`2021-07-24 16:36:45,713 - mmdet - INFO - Epoch [1][100/14009] lr: 4.653e-05, eta: 13 …
-
It seems that the model will converge much more slower if **wh_offset_base =1**, especially the wh_loss. Besides, compared with centernet, the loss weight of wh_loss is much more larger than hm_loss（c…
-
Thanks for sharing your work.
I have one question: in tab 1 of the paper, if not use wh_gaussian, the accuracy reduce from 32.9 to 32. I want to reproduce the result of 32. If set wh_gaussian=False, …
-
在使用不同模型的时候发现了一个问题，由于我自己的数据集的json文件中类别的序号是从0开始的，在跑一些实验的时候是正常可行的，例如TTFNet、PAFNet；但是我今天在进行GFLv2的实验时则发现，GFLv2的类别会自动增加一类背景类，因而在进行eval的过程中会导致类别索引为6但是val.json中最大索引只到5(数据集0-5共6个类别)。
因此该如何修改GFLv2的代码使得其适应当前数据集…
ghost updated
2 years ago
-
@liuzili97 @Turoad
**Motivation**
Update current codes into the latest mmdetection is helpful.
**Additional context**
By simply copying the model and config files and modifying import settings…
-
尊敬的开发者，你好，
请问一下，[cornernet_squeeze_hg104.yml](https://github.com/PaddlePaddle/PaddleDetection/blob/master/configs/anchor_free/cornernet_squeeze_hg104.yml)里面是只用了Soft-NMS和detections_per_im这两种后处理操作吗？
h…
-
### 问题确认 Search before asking
- [X] 我已经查询[历史issue](https://github.com/PaddlePaddle/PaddleDetection/issues)，没有发现相似的bug。I have searched the [issues](https://github.com/PaddlePaddle/PaddleDetection/issu…
-
# Paddle2ONNX v0.8 (2021.08)
- [ ] Demo shows how to deploy Paddle model use C# through onnxruntime
- [ ] Quantize Model Support
# Paddle2ONNX v0.7 (2021.07)
- [ ] OpenVINO/OpenCV inference supp…
-
**CenterNet**
[Objects as Points](https://arxiv.org/abs/1904.07850) seems to achieve a good speed-accuracy tradeoff, better than Yolo v3, and probably better than CornerNet (#3229).
GitHub repo…
-
作者您好，您提供的ROMP_HRNet32_V1.pkl是否是通过v1.yml在pretrain_hrnet.pkl的基础上训练得到的吗？