UnboundLocalError: local variable 'outs' referenced before assignment

[BillSJC]

caught an err when using the smple code

UnboundLocalError: local variable 'outs' referenced before assignment

  File "C:\Python36\lib\site-packages\tensorflow\python\keras\engine\training_generator.py", line 358, in evaluate_generator
    if not isinstance(outs, list):
UnboundLocalError: local variable 'outs' referenced before assignment

from imageai.Prediction.Custom import ModelTraining
model_trainer = ModelTraining()
model_trainer.setModelTypeAsResNet()
model_trainer.setDataDirectory("pets")
model_trainer.trainModel(num_objects=2, num_experiments=2, enhance_data=True, batch_size=4, show_network_summary=True)

and here are the struct of folder pets

[OlafenwaMoses]

How many pictures are in the test and train folder for each object class?

[BillSJC]

i do it only for test so i use a small amount of pictures

test:

• cat:1
• dog:2

train:

• cat:7
• dog:9

[OlafenwaMoses]

The recommended number of training images for each object class is 500. Check out the documentation link below.

https://imageai.readthedocs.io/en/latest/custom/index.html