训练问题

[justcodew]

使用作者放在百度网盘的模型作为初始模型，数据使用CCPD
数据格式为： 类别、中心坐标、中心宽高、四个角点 （做了归一化）
训练效果比较奇怪，从55到56epoch精度和召回率直接下降到0

[justcodew]

数据格式
3432-20_6-47&274_673&732-673&516_110&732_47&490_610&274-0_0_2_11_33_28_30-108-396
box : 47&274_673&732
points : 673&516_110&732_47&490

转换为
0 0.5 0.4336206896551724 0.8694444444444445 0.39482758620689656 0.9347222222222222 0.44482758620689655 0.1527777777777778 0.6310344827586207 0.06527777777777778 0.4224137931034483 0.8472222222222222 0.23620689655172414
从训练日志中可视化的图片看 数据标签应该是没有问题的

[ouxiand]

请问你解决了吗？ 我怎么训练都有bug 处理一个又一个

[ouxiand]

数据格式和你是一样的

[ouxiand]

Traceback (most recent call last):
File "/detect/yolov5-car-plate/./train.py", line 552, in
train(hyp, opt, device, tb_writer)
File "/detect/yolov5-car-plate/./train.py", line 285, in train
for i, (imgs, targets, paths, _) in pbar: # batch -------------------------------------------------------------
File "/root/anaconda3/lib/python3.9/site-packages/tqdm/std.py", line 1195, in iter
for obj in iterable:
File "/detect/yolov5-car-plate/utils/plate_datasets.py", line 104, in iter
yield next(self.iterator)
File "/root/anaconda3/lib/python3.9/site-packages/torch/utils/data/dataloader.py", line 521, in next
data = self._next_data()
File "/root/anaconda3/lib/python3.9/site-packages/torch/utils/data/dataloader.py", line 1183, in _next_data
return self._process_data(data)
File "/root/anaconda3/lib/python3.9/site-packages/torch/utils/data/dataloader.py", line 1229, in _process_data
data.reraise()
File "/root/anaconda3/lib/python3.9/site-packages/torch/_utils.py", line 434, in reraise
raise exception
TypeError: Caught TypeError in DataLoader worker process 1.
Original Traceback (most recent call last):
File "/root/anaconda3/lib/python3.9/site-packages/torch/utils/data/_utils/worker.py", line 287, in _worker_loop
data = fetcher.fetch(index)
File "/root/anaconda3/lib/python3.9/site-packages/torch/utils/data/_utils/fetch.py", line 49, in fetch
data = [self.dataset[idx] for idx in possibly_batched_index]
File "/root/anaconda3/lib/python3.9/site-packages/torch/utils/data/_utils/fetch.py", line 49, in
data = [self.dataset[idx] for idx in possibly_batched_index]
File "/detect/yolov5-car-plate/utils/plate_datasets.py", line 530, in getitem
img, labels = load_mosaic(self, index)
File "/detect/yolov5-car-plate/utils/plate_datasets.py", line 758, in load_mosaic
x,y = point
TypeError: cannot unpack non-iterable numpy.float32 object

[ouxiand]

数据集
0 0.4346698113207547 0.5932914046121593 0.2761399371069182 0.2865129280223619 0.33885613207547166 0.6457023060796646 0.40469732704402517 0.6561844863731656 0.40469732704402517 0.6876310272536688 0.34180424528301884 0.6771488469601677

[ouxiand]

数据集
0 0.4346698113207547 0.5932914046121593 0.2761399371069182 0.2865129280223619 0.33885613207547166 0.6457023060796646 0.40469732704402517 0.6561844863731656 0.40469732704402517 0.6876310272536688 0.34180424528301884 0.6771488469601677

[justcodew]

没解决 使用了其他旋转目标检测的方法
https://github.com/open-mmlab/mmrotate