ERROR with shape size when evaluating custom dataset model

[francismontalbo]

Hello everyone,

I trained my model then when I was about to evaluate/inference it I get this error.

ValueError: You are trying to load a weight file containing 132 layers into a model with 131 layers.

I trained the model well with no problems. However, when I tried to use the inference it is giving me this error.

[francismontalbo]

I fixed it but this error is coming up when I evaluate

ValueError: shapes (2,262144) and (3136,1) not aligned: 262144 (dim 1) != 3136 (dim 0)

[vkramdev]

Hi @francismontalbo , were you able to resolve the issue?

I am facing the same issue when trying to run utils.compute_ap.

Getting error: ValueError: shapes (1,1048576) and (3136,1) not aligned: 1048576 (dim 1) != 3136 (dim 0)

I have trained my model on one object class.

[francismontalbo]

Include the mini mask into your config then set it to false :) @vkramdev

[sdaemi]

hi francis, how did you sort the first issue? getting a similar error:

ValueError: You are trying to load a weight file containing 234 layers into a model with 233 layers.

I am running tensorflow 2.20

[raficabral]

I am facing the same issue @francismontalbo @vkramdev . Did u fix this and could help me?

[tanuj1221]

Include the mini mask into your config then set it to false :) @vkramdev

but it is giving 'load_image_gt() got an unexpected keyword argument 'use_mini_mask'

[joseveloso1501]

@tanuj1221 because 'use_mini_mask' is not an argument of 'load_image_gt()', add it into your Config Class