jolt2017

Results 11 issues of jolt2017

Recently i want to genetate some celeba_hq image.but the code lacks detail comments on the part of creat_celeba_hq funtcion. If comments are possible, that must be a gift for me...

hi: I confronted some problems when run vggface = VGGFace(include_top = False, model = 'resnet50', input_shape = (224, 224, 3)) Cannot interpret feed_dict key as Tensor: Tensor Tensor("Placeholder_798:0", shape=(7, 7,...

hi: I confronted some problems when run vggface = VGGFace(include_top = False, model = 'resnet50', input_shape = (224, 224, 3)) ![image](https://user-images.githubusercontent.com/45425532/64471773-20596e00-d188-11e9-9134-f499d4fa0c51.png) Did you have any suggestions to solve this problem?

您好,请问 solov2 训练好模型以后,如何计算他在测试集上 分割精度指标能?

![image](https://user-images.githubusercontent.com/45425532/93443786-f9895800-f902-11ea-97cf-06846da7b640.png) In your paper, you said your discriminator is a kind of multisacle discriminator, but in your code, line 240 and line 241 in the picture, I noted that self.scale...

Great work. So do u have the schedule to realease the traing code?

考虑 ncnn 部署代码的实现么

hi: 支持一波,写的很棒,已经用上了,感谢。发现一个小问题,训练时 改变 train.sh gpus 参数 并不能指定 gpu。

请问最新发布的 paddleHumanSeg 里面的 SPPM 模块在哪个脚本里? https://mp.weixin.qq.com/s/XkLJU7CvyGs7NCF0ypehFg

question

你好,查看 tps 相关代码,发现有个接口不是很理解,截图放在下面的网盘链接上了,可否告知输入及返回参数的含义. https://github.com/PaddlePaddle/PaddleOCR/blob/release/2.6/ppocr/modeling/transforms/tps.py 链接:https://pan.baidu.com/s/1wAfkkSvnuFF-k3kBPAWinA?pwd=1bbx 提取码:1bbx