CODEWITHSUNDEEP
keras
EhsanYaghoubi
EhsanYaghoubi

Reputation: 145

TypeError: load_model() missing 1 required positional argument: 'filepath'

I am using Keras 2.2.5 in the training phase. I have saved the model with ModelCheckpoint function that is imported like this: from keras.callbacks import ModelCheckpoint

Then, in the test phase, when I want to load the model using the load_model function (from keras.models import load_model), I get the title error.

test script is as follow:

import numpy as np
import argparse
import keras.layers as KL
from keras.models import load_model
from keras.preprocessing import image
from keras.applications.resnet50 import preprocess_input as preprocess_input_resnet
if __name__ == "__main__":
    # Define variables
    parser = argparse.ArgumentParser()
    parser.add_argument("--image_size", type=tuple, default=(500, 500))
    parser.add_argument("--mask_size", type=tuple, default=(32,32))
    parser.add_argument("--image_path", type=str, default="../DATA/resized_imgs/13056.png")
    parser.add_argument("--mask_path", type=str, default="../DATA/resized_masks/13056.png")
    parser.add_argument("--path_of_the_checkpoint", type=str, default="./RESULTS/2020_02_25_12_34_54/bestmodel/MultiLabel_PETA_weights.best.hdf5")
    parser.add_argument("--Categories", type=list, default=["personalLess30", "personalLess45", "personalLess60", "personalLarger60"])
    args = parser.parse_args()

    # Load trained model
    PAR_model = load_model (filepath=args.path_of_the_checkpoint, custom_objects={'BatchNorm':KL.BatchNormalization})
    PAR_model.summary()

Upvotes: 1

Views: 1807

Answers (1)

Dr. Snoopy
Dr. Snoopy

Reputation: 56377

You should not pass positional arguments as keyword arguments:

PAR_model = load_model (args.path_of_the_checkpoint, custom_objects={'BatchNorm':KL.BatchNormalization})

Upvotes: 2

Related Questions