 RenLiqiang
|
a59b471b64
修改LineDetect为keypoint方式
|
5 months ago |
|
RenLiqiang
|
eb72a21f32
add lines_generator
|
5 months ago |
 lstrlq
|
0c68613b41
调试训练代码参数为512分辨率
|
6 months ago |
|
RenLiqiang
|
8a2cadb18a
调试maskrcnn 训练螺纹钢
|
7 months ago |
|
RenLiqiang
|
28658aaa06
重构Trainer
|
7 months ago |
|
RenLiqiang
|
3b8e1a3d58
添加记录验证机损失功能
|
7 months ago |
|
RenLiqiang
|
34fddfcafe
添加根据val_loss保存权重功能
|
7 months ago |
|
lstrlq
|
ebb3401c24
添加深度值归一化功能
|
7 months ago |
|
RenLiqiang
|
34d748b2dc
修改roihead
|
7 months ago |
|
RenLiqiang
|
7ccb211c2a
修改删除根据点的种类进行match
|
7 months ago |
|
RenLiqiang
|
2c1a9b7db1
tensorboard 加入lmap jmap
|
8 months ago |
 xue50
|
45170b88b6
不使用line
|
8 months ago |
|
xue50
|
c2cffb6b1a
predict_box内无线段时，选box内点组成线段最长的 两个点组成的线段返回
|
8 months ago |
|
RenLiqiang
|
d986c2687c
修改线段采样函数与前向传播采样方法不一致导致推理线段效果不好的bug
|
8 months ago |
|
xue50
|
d0d3f9387f
OK
|
10 months ago |
|
xue50
|
8c208a87b7
OK
|
10 months ago |
|
xue50
|
6251ff34cd
self.training is wrong
|
10 months ago |
|
xue50
|
6333887184
Line_rcnn and ROI_head
|
10 months ago |
