pytorch_solov2 icon indicating copy to clipboard operation
pytorch_solov2 copied to clipboard

SOLOv2 pytorch (without mmdetection and mmcv)

Results 16 pytorch_solov2 issues
Sort by recently updated
recently updated
newest added

Traceback (most recent call last): File "G:/pycharmProject/pytorch_solov2-master/train.py", line 4, in from modules.solov2 import SOLOV2 File "G:\pycharmProject\pytorch_solov2-master\modules\solov2.py", line 6, in from .solov2_head import SOLOv2Head File "G:\pycharmProject\pytorch_solov2-master\modules\solov2_head.py", line 8, in from .focal_loss...

在'solov2_head.py'里,'solov2_target_single'函数会生成ground truth label, cate_label初始值是0: cate_label = torch.zeros([num_grid, num_grid], dtype=torch.int64, device=device) 然后在下边根据中心点的位置,更新cate_label: cate_label[top:(down+1), left:(right+1)] = gt_label 但class是从0-80,第0类是人,物体很多时候并不是布满整幅图的,这样会cate label里就有大量的0类 有人可以帮忙解释一下不?谢谢

RuntimeError: one of the variables needed for gradient computation has been modified by an inplace operation: [torch.FloatTensor [1, 128, 192, 120]], which is output 0 of ReluBackward 0, is at...

When running ``` python setup.py develop on colab ``` Ran into the following error ``` running develop running egg_info creating focalloss.egg-info writing focalloss.egg-info/PKG-INFO writing dependency_links to focalloss.egg-info/dependency_links.txt writing top-level names...

Hi, I am gonna converting SOLOv2 pytorch model to onnx? You can tell me the solution? Thank you :D

@ABlueLight 请问您和官方对比过主干网络分别是resnet18和34的精度嘛?

我训练初始损失3+最后损失值降低到了0.5,但是利用提供的可视化代码显示效果并不理想.