ImageAI/Keras can't load ResNet model

Question

I'm trying to use the Image Prediction tool in the imageai library, and I'm getting the following two errors... As a beginner, I honestly can't make sense of the errors, and I couldn't find any answers online that worked. If anyone could please help me solve it I would greatly appreciate it.

1.

ImportError: load_weights requires h5py when loading weights from HDF5.

But I do have h5py installed and updated. I also tried installing h5py==2.10.0 and cython, as others suggested in past questions, but that hasn't worked for me either.

2.

ValueError: You have specified an incorrect path to the ResNet model file.

I have tried several different ways of writing the path but this still won't work.

This is the full error text:

Traceback (most recent call last):
  File "C:\Users\MYUSER\Miniconda3\envs\tensorflow\lib\site-packages\imageai\Prediction\__init__.py", line 125, in loadModel
    model = ResNet50(model_path=self.modelPath, model_input=image_input)
  File "C:\Users\MYUSER\Miniconda3\envs\tensorflow\lib\site-packages\imageai\Prediction\ResNet\resnet50.py", line 115, in ResNet50
    model.load_weights(weights_path)
  File "C:\Users\MYUSER\Miniconda3\envs\tensorflow\lib\site-packages\tensorflow\python\keras\engine\training.py", line 2341, in load_weights
    raise ImportError(
ImportError: `load_weights` requires h5py when loading weights from HDF5.

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "C:\Users\MYUSER\Current_Working_Directory\brain.py", line 10, in <module>
    prediction.loadModel() 
  File "C:\Users\MYUSER\Miniconda3\envs\tensorflow\lib\site-packages\imageai\Prediction\__init__.py", line 129, in loadModel
    raise ValueError("You have specified an incorrect path to the ResNet model file.")
ValueError: You have specified an incorrect path to the ResNet model file.

This is my code:

from imageai.Prediction import ImagePrediction 
import os 
execution_path=os.getcwd() 

prediction = ImagePrediction()
prediction.setModelTypeAsResNet() 
prediction.setModelPath(os.path.join(execution_path, "resnet50_imagenet_tf.2.0.h5")) 
prediction.loadModel() 

predictions, probabilities = prediction.predictImage(os.path.join(execution_path, "giraffe.jpg"), result_count=5 )

for eachPrediction, eachProbability in zip(predictions, probabilities): 
    print(eachPrediction , " : " , eachProbability)

score 0 · Answer 1 · answered Jul 19 '22 at 14:25

Try this:

from imageai.Detection import ObjectDetection
import os

execution_path = os.getcwd()

detector = ObjectDetection()
detector.setModelTypeAsRetinaNet()
detector.setModelPath( os.path.join(execution_path , "resnet50_coco_best_v2.1.0.h5"))
detector.loadModel()

detections, objects_path = detector.detectObjectsFromImage(input_image=os.path.join(execution_path , "input.jpg"), output_image_path=os.path.join(execution_path , "output.jpg"), minimum_percentage_probability=10,  extract_detected_objects=True)

for eachObject, eachObjectPath in zip(detections, objects_path):
    print(eachObject["name"] , " : " , eachObject["percentage_probability"], " : ", eachObject["box_points"] )
    print("Object's image saved in " + eachObjectPath)
    print("--------------------------------")

The Imagenet-algorithm didnt work for me either, but using the coco-algorithm (download on Github of ImageAI) saves me a lot of time and effort atm.

Also your code is not suited for windows, since it requires the direct pathings to the files, otherwise you will get that exact error.

ImageAI/Keras can't load ResNet model

1 Answers1