search

PaddlePaddle / PaddleDetection

Object Detection toolkit based on PaddlePaddle. It supports object detection, instance segmentation, multiple object tracking and real-time multi-person keypoint detection.
Apache License 2.0
12.65k stars 2.87k forks source link

labelImg标注完的数据怎么数据增强进入模型的训练？ #4896

Open LMR2018 opened 2 years ago

LMR2018 commented 2 years ago

labelImg标注完的数据怎么数据增强进入模型的训练？

wangxinxin08 commented 2 years ago

数据增强是在Reader里面配置的，比如https://github.com/PaddlePaddle/PaddleDetection/blob/release/2.3/configs/ppyolo/base/ppyolov2_reader.yml#L5 这里的smaple transform和batch transform都是数据增强的方法，直接配置即可调用

LMR2018 commented 2 years ago

数据增强是在Reader里面配置的，比如https://github.com/PaddlePaddle/PaddleDetection/blob/release/2.3/configs/ppyolo/base/ppyolov2_reader.yml#L5 这里的smaple transform和batch transform都是数据增强的方法，直接配置即可调用

@wangxinxin08 你训练配置上尝试过这个参数吗？AutoAugment: {autoaug_type: v1}，我的使用这个配置会报错，[12/18 17:13:32] reader WARNING: fail to map sample transform [AutoAugment_f7f412] with error: Cannot handle this data type: (1, 1, 4), <f4 and stack: Traceback (most recent call last): File "C:\Users\Dev16\Anaconda3\lib\site-packages\PIL\Image.py", line 2680, in fromarray mode, rawmode = _fromarray_typemap[typekey] KeyError: ((1, 1, 4), '<f4')