Wizyoung

周末我尝试编译一下吧

突然想起这件事... 3.8 的版本在我的这条[评论](https://github.com/pytorch/pytorch/issues/48145#issuecomment-747631341) 后面下一条已经有人跟进编译了 3.8 的版本, 地址是 (Google Drive) [torch-1.8.0a0-cp38-cp38-macosx_11_0_arm64.whl](https://drive.google.com/file/d/1e-7R3tfyJqv0P4ijZOLDYOleAJ0JrGyJ/view?usp=sharing) 试试看能不能用? 不行的话我就给你再编译一个出来

怎么可能是验证集和训练集一样呢。。。。。。 参数什么的也要自己调的啊，我只是放上了我使用的默认的参数。多调多观察吧

Hi, I guess in your TFlite version, ResizeNearestNeighbor of tf.image module has not been supported. According to this [Link](https://github.com/tensorflow/tensorflow/issues/23937), you can try newer tflite version. Or you may change this...

@yuanzhedong Hi, you can refer to my answer in stackoverflow https://stackoverflow.com/a/49768479/6631854 to get the correct placeholder name.

Results on epoch 99 is kind of overfitting. I got 87.54% mAP after 36 epochs. Here is my training logs for your reference: [training.log](https://github.com/wizyoung/YOLOv3_TensorFlow/files/3389002/training.log)

@liminghuiv Are your training and test txt files correct? Here are my txt files: [train.txt](https://github.com/wizyoung/YOLOv3_TensorFlow/files/3391064/train.txt) [val.txt](https://github.com/wizyoung/YOLOv3_TensorFlow/files/3391065/val.txt) I hope you make efforts to understand the yolo v3 model and its parameters,...

I think this makes no sense. Questions are like, how to choose prediction scale using IoU ranking, or how to set ignore mask. Of course you can modify the code...

I will **not** answer questions without detailed descriptions like error info.

@nobody-cheng No, it's consistent with the original paper. Note that the formula you cited is a mask array. The log part is in the conf_loss part and conf_loss *= focal_mask....