Custom model train gives error: You are trying to load a weight file containing 116 layers into a model with 106 layers

[sb3114]

hi,
i am new to learning AI/ML using imageai, tensorflow etc. I am testing custom training using transfer learning as i have 5 images only ( 3 for train and 2 for test), i get below error. what does this mean really?

below is my system settings:
raspberry Pi OS 64 bullseye
Keras 2.4.3
Keras-Preprocessing 1.1.2
python 3.7.7

My code - same error with resnet50_coco_best_v2.1.0.h5 , resnet50_imagenet_tf.2.0.h5:

(virtenv3.7) pi@pi:~/ml/detection $ cat customtrain.py
from imageai.Classification.Custom import ClassificationModelTrainer
model_trainer = ClassificationModelTrainer()
model_trainer.setModelTypeAsResNet50()
model_trainer.setDataDirectory("mask")
model_trainer.trainModel(num_objects=2, num_experiments=10, enhance_data=True, batch_size=2, show_network_summary=True, transfer_from_model="resnet50_coco_best_v2.0.1.h5", initial_num_objects=1000)
(virtenv3.7) pi@pi:~/ml/detection $

Error:
(virtenv3.7) pi@pi:~/ml/detection $ python customtrain.py Traceback (most recent call last): File "customtrain.py", line 5, in <module> model_trainer.trainModel(num_objects=2, num_experiments=10, enhance_data=True, batch_size=2, show_network_summary=True, transfer_from_model="resnet50_coco_best_v2.0.1.h5", initial_num_objects=1000) File "/home/pi/virtenv3.7/lib/python3.7/site-packages/imageai/Classification/Custom/__init__.py", line 225, in trainModel include_top=False, pooling="avg") File "/home/pi/virtenv3.7/lib/python3.7/site-packages/tensorflow/python/keras/applications/resnet.py", line 472, in ResNet50 input_tensor, input_shape, pooling, classes, **kwargs) File "/home/pi/virtenv3.7/lib/python3.7/site-packages/tensorflow/python/keras/applications/resnet.py", line 221, in ResNet model.load_weights(weights) File "/home/pi/virtenv3.7/lib/python3.7/site-packages/tensorflow/python/keras/engine/training.py", line 2234, in load_weights hdf5_format.load_weights_from_hdf5_group(f, self.layers) File "/home/pi/virtenv3.7/lib/python3.7/site-packages/tensorflow/python/keras/saving/hdf5_format.py", line 688, in load_weights_from_hdf5_group ' layers.') ValueError: You are trying to load a weight file containing 116 layers into a model with 106 layers. (virtenv3.7) pi@pi:~/ml/detection $

[sb3114]

changed to using DetectionModelTrainer instead of Classification and used yolo model which solved the issue. Not sure why
`from imageai.Detection.Custom import DetectionModelTrainer

trainer = DetectionModelTrainer()
trainer.setModelTypeAsYOLOv3()
trainer.setDataDirectory(data_directory="./imageai/data/")
trainer.setTrainConfig(object_names_array=["with_mask", "without_mask","mask_weared_incorrect"],
batch_size=8,
num_experiments=10,
train_from_pretrained_model="pretrained-yolov3.h5")

trainer.trainModel()
`