Question

我正在尝试使用GitHub上的YunYang1994 / tensorflow-yolov3实现并使用YoloV3训练我的自定义数据：http://www.pydev.org/manual_101_navigate_code.html

我的数据有2个类别：1类有80张火车图像，20张测试图像。第2类有48列火车，有12项测试。

我准备了所需的一切： train.tfrecords，test.tfrecords ...

但是当我运行quick_train（修改路径之后）时，我得到了：

Traceback (most recent call last):
  File "thao_train.py", line 47, in <module>
    loss             = model.compute_loss(pred_feature_map, y_true)
  File "C:\Users\thao\tensorflow-yolov3\core\yolov3.py", line 263, in compute_loss
    result = self.loss_layer(pred_feature_map[i], y_true[i], _ANCHORS[i])
  File "C:\Users\thao\tensorflow-yolov3\core\yolov3.py", line 318, in loss_layer
    true_tw_th = y_true[..., 2:4] / anchors
  File "C:\ProgramData\Anaconda3\lib\site-packages\tensorflow\python\ops\math_ops.py", line 824, in binary_op_wrapper
    return func(x, y, name=name)
  File "C:\ProgramData\Anaconda3\lib\site-packages\tensorflow\python\ops\math_ops.py", line 920, in _truediv_python3
    return gen_math_ops.real_div(x, y, name=name)
  File "C:\ProgramData\Anaconda3\lib\site-packages\tensorflow\python\ops\gen_math_ops.py", line 6897, in real_div
    "RealDiv", x=x, y=y, name=name)
  File "C:\ProgramData\Anaconda3\lib\site-packages\tensorflow\python\framework\op_def_library.py", line 788, in _apply_op_helper
    op_def=op_def)
  File "C:\ProgramData\Anaconda3\lib\site-packages\tensorflow\python\util\deprecation.py", line 507, in new_func
    return func(*args, **kwargs)
  File "C:\ProgramData\Anaconda3\lib\site-packages\tensorflow\python\framework\ops.py", line 3300, in create_op
    op_def=op_def)
  File "C:\ProgramData\Anaconda3\lib\site-packages\tensorflow\python\framework\ops.py", line 1823, in __init__
    control_input_ops)
  File "C:\ProgramData\Anaconda3\lib\site-packages\tensorflow\python\framework\ops.py", line 1662, in _create_c_op
    raise ValueError(str(e))
ValueError: Dimensions must be equal, but are 3 and 0 for 'yolov3/truediv_8' (op: 'RealDiv') with input shapes: [?,13,13,3,2], [0,2].

有人猜为什么会这样吗？

在YOLOv3 YunYang1994实现中训练客户数据时，如何解决ValueError？

0 个答案: