yolov5-face

yolov5-face copied to clipboard

May I ask if yolo5face did not use classification loss for reverse rafter propagation

xailxail

yolo8face 何时出

2

666tua

RuntimeError: Expected all tensors to be on the same device, but found at least two devices, cuda:0 and cpu! (when checking argument weight in method wrapper__cudnn_convolution) 我已经将数据和模型.to(device)，依然报错，报错部分为： 

WisonZ

Hi, I was interested in your stem block, so I added the stem block module in the ultralytics yolov5 training pipeline. All went fine when depth multiple = 0.33 and...

junghyun-avikus

lks_mask = torch.where(lks < 0, torch.full_like(lks, 0.), torch.full_like(lks, 1.0)) 代码不应该再下列代码之后？ #应该是关键点的坐标除以anch的宽高才对，便于模型学习。使用gwh会导致不同关键点的编码不同，没有统一的参考标准 lks[:, [0, 1]] = (lks[:, [0, 1]] - gij) lks[:, [2, 3]] = (lks[:, [2, 3]] - gij) lks[:,...

cqray1990

@derronqi I've found possible errors in face_datasets.py. https://github.com/deepcam-cn/yolov5-face/blob/3d1cb7ab282dc00b7c9ff0ab4335e24ece7d1bed/utils/face_datasets.py#L317 https://github.com/deepcam-cn/yolov5-face/blob/3d1cb7ab282dc00b7c9ff0ab4335e24ece7d1bed/utils/face_datasets.py#L319 I think it should be `x[:, 9] > 0` and `x[:, 10] > 0` in the above code, respectively, as the...

rcg12387

Hello, I would like to ask a few questions about this github repo. 1. What do I need to do to train using the pretrained model? 2. How can I...

SwEngine

低分辨率下landmark值不准确

1

在320分辨率下推理，landmark的值不准确，会存在一定的偏差，在640分辨率下是相对准确的。尝试过320和640分辨率的训练，训练集都是widerface，结果显示都是如此。怀疑过是C++代码的问题，用python预测pt模型，也存在偏差。 大家有碰到这种情况吗？

408550969

为什么验证集的标签格式和训练集的标签格式不一样？

1

请问val中的标签格式后10位数字也就是五个关键点的坐标值都是-1，而train中的labels确是完整的，这是什么原因？

1234dgfg

cqray1990