CornerNet-Lite
CornerNet-Lite copied to clipboard
princeton-vl
Hi, I am running CornerNet- Squeeze on real time video using the code below. But the result is quite bad in terms of inferencing speed. It is almost 1s/ frame....
Why real-time detection cornernet-lite does not work as well as YOLOV3
when i run evaluate.py, I get some trouble like this: Traceback (most recent call last): File "evaluate.py", line 104, in main(args) File "evaluate.py", line 100, in main test(testing_db, system_config, model,...
When I train my own data, used CornerNet_saccade model, but the loss do not decrease, in my training process. I do not known why, should I need a pre_trained model?
Hi, author! I wanna ask some questions about the training speed of your CornerNet_Squeeze with the default hyper-parameters in your configuration. How long did you take to get the results...
我是按照作者的目录放置的文件，代码也不会有错，但是运行的时候出现了“None type has attribute shape”后来同事提醒发现，应该把minival2014里面的图片全部复制然后放到trainval2014里面合并在一块。然后就训练成功跑通了。
hi, after there a some issues related to images not loading, i suggest to add some error handling after: https://github.com/princeton-vl/CornerNet-Lite/blob/6a54505d830a9d6afe26e99f0864b5d06d0bbbaf/core/sample/cornernet_saccade.py#L162 I would add after `image = cv2.imread(image_path)`: ```python if image...
[  ](url) [  ](url) it is clear that the bound boxes' size are wrongly identified for that the cornerpoints are wrongly grouped. is there any ways such as...
  First Image is the CornerNet-Squeeze result , second Image is yolo v3 result
