Detection using custom model

[John-Almardeny]

I am getting this error:

Traceback (most recent call last):
  File "/home/yahya/Desktop/ML/FirstDetection.py", line 11, in <module>
    detector.loadModel()
  File "/usr/local/lib/python3.5/dist-packages/imageai/Detection/__init__.py", line 185, in loadModel
    model.load_weights(self.modelPath)
  File "/usr/local/lib/python3.5/dist-packages/keras/engine/network.py", line 1166, in load_weights
    f, self.layers, reshape=reshape)
  File "/usr/local/lib/python3.5/dist-packages/keras/engine/saving.py", line 1030, in load_weights_from_hdf5_group
    str(len(filtered_layers)) + ' layers.')
ValueError: You are trying to load a weight file containing 107 layers into a model with 116 layers.

My code is as simple as this:

from imageai.Detection import ObjectDetection
import os

execution_path = os.getcwd()
detector = ObjectDetection()
detector.setModelTypeAsRetinaNet()
detector.setModelPath(os.path.join(execution_path , "custom_model.h5"))
detector.loadModel()
detections = detector.detectObjectsFromImage(input_image=os.path.join(execution_path , "2.png"))

[OlafenwaMoses]

Hi John, you are trying to perform object detection using the prediction model you trained. Please be aware your custom model is for prediction and not object detection. Support for training custom models for object detection will be provided in subsequent versions of ImageAI.

Also, ensure you set the model type of your prediction class to correspond with the model type you used in straining.

[John-Almardeny]

Thanks. That makes sense now.

[rw86347]

I just realized this too. Do you have a time frame for when detection will be supported?

[OlafenwaMoses]

Please refer to #8