Value error

[Dev0104]

I am getting this error when trying to train the model.

'ValueError: Dimensions must be equal, but are 37 and 18 for '{​​​​​​​​{​​​​​​​​node rpn_loss_cls_fixed_num/mul}​​​​​​​​}​​​​​​​​ = Mul[T=DT_FLOAT](rpn_loss_cls_fixed_num/strided_slice_1, rpn_loss_cls_fixed_num/Log)' with input shapes: [1,37,50,9], [1,18,25,9].'

Is it something related to the input and output shapes or anchors?
Can someone please help?

[NaisargiOza]

Hi,
I am getting the same error. Were you able to resolve this error?

[HebdaMarta]

I had the same problem and I changed last line in nn_base model for:
x = MaxPooling2D((1, 1), strides=(1, 1), name='block5_pool')(x)

You need to adjust the size of your CNN