libfacedetection.train
libfacedetection.train copied to clipboard
ShiqiYu
The training program for libfacedetection for face detection and 5-landmark detection.
模型为640x640，但是为转onnx的时候尺寸改为641的大小也一样可以转onnx，请问这是为什么呢？不是规定了是640x640的大小吗？
Currently, I have some problems converting the model trained by this repo, but successfully convert the model at the OpenCV GitHub [repo](https://github.com/opencv/opencv_zoo/tree/master/models/face_detection_yunet). So I wonder if the training code is...
于老师,您好 我在使用项目中提供的训练脚本得到的模型性能比您开源的模型性能要差1~1.5个点.请问开源模型是否在训练时进行了别的参数调优,还是仅仅是因为数据shuffle不同导致的? 此外,还想请教一下,如果我们不需要输出人脸关键点预测,那么在训练时删除人脸关键点,对于人脸检测性能会有帮助吗?
//only 27 elements used for each pixel create((imgHeight+1)/2, (imgWidth+1)/2, 32); //since the pixel assignment cannot fill all the elements in the blob. //some elements in the blob should be initialized...
请问余老师ConvDPUnit中的卷积部分如下代码：  为什么没有使用pointwise的操作进行降维呢？这样操作对速度和精度有什么作用吗？谢谢
As i am testing the model, i encounter it creates a lot of false positive detection on hands and necks with high confidence, can you tell me how to add...
您好，目前代码是支持训练固定次迭代，没有val集做early-stop个人认为会难以捕捉over-fitting的情况，请问后续有添加eval.py的计划吗？我自己写了一份eval.py，但是不知道对不对。 此致敬礼，顺颂时祺
Hi, I manually label custom datasets following the COCO format. May I ask if is it possible for you to publish the code to convert to NVIDIA Dali (which is...
带出onnx出错
使用命令： `python tools/yunet2onnx.py ./configs/yunet_n.py ./weights/yunet_n.pth` 导出模型时出现这个错误： ``` File "tools/yunet2onnx.py", line 265, in dynamic_export=args.dynamic_export) File "tools/yunet2onnx.py", line 96, in pytorch2onnx dynamic_axes=dynamic_axes) File "C:\Users\jiafc\AppData\Local\conda\conda\envs\yoloAir\lib\site-packages\torch\onnx\__init__.py", line 365, in export export_modules_as_functions, File "C:\Users\jiafc\AppData\Local\conda\conda\envs\yoloAir\lib\site-packages\torch\onnx\utils.py", line...
