An error occurred: No file or directory found at models-kp/model_11

Question

I want to run a ready-trained model I found on github. This model is about image processing and scoring via images. In the introduction it says it will work if you do this: python3 main.py path/to/your/image.

This is github link: https://github.com/dominika01/pain-detection-cats

I adapted my path to this code and ran it on cmd and terminal (I've checked from 2 os for error).

Loading the image… Cropping the features… Loading model… Issue loading image

There are multiple .py files and files for more than one model in the folder. I opened main.py with spyder. ` import cv2 import numpy as np import tensorflow as tf import keypoints import joblib import sys

 def preprocess_feature (feature, x_size, y_size):
 resized = cv2.resize(feature, (x_size, y_size))
 preprocessed = tf.keras.preprocessing.image.img_to_array(resized)
 reshaped = preprocessed.reshape(1, x_size, y_size, 3)
 return reshaped

 def score_feature (feature, model):
 probability_array = model.predict(feature)
 score = np.argmax(probability_array, axis=1)
 score = score[0]
 confidence = round(np.max(probability_array), 2)
 if confidence > 0.99:
     score = 0
 return score

# path to the image
print("Loading the image...")
#image_path = PATH
image_path = sys.argv[1]

`

This time too Loading the image… Cropping the features… Loading model… Issue loading image

Here sys.argv[1] causes the "IndexError: list index out of range" error. For that reason I changed it to image_path = ("/Users////.jpg"). Again: Loading the image… Cropping the features… Loading model… Issue loading image I'm stuck in the phase.

Afterwards I made a few corrections. Error: File format not supported: filepath=models-kp/model_11. Keras 3 only supports V3 .keras files and legacy H5 format files (.h5 extension). Note that the legacy SavedModel format is not supported by load_model() in Keras 3. In order to reload a TensorFlow SavedModel as an inference-only layer in Keras 3, use keras.layers.TFSMLayer(models-kp/model_11, call_endpoint=' serving_default') (note that your call_endpoint may have a different name).

I got an error like this. Afterwards, I downgraded my Keras and TF version. right now just I get the error An error occurred: No file or directory found at models-kp/model_11. Yes, there is no folder named model-kp, there is no model_1, 2 …11 in this GitHub folder. This is in keypoints.py

   # global variables
   INPUT_SHAPE = 256
   **ITERATION = '11'**
   EPOCHS = 32
   BATCH_SIZE = 32

    def setup_paths():
    folder_dir = "data-keypoints"
    folders = ['CAT_00', 'CAT_01', 'CAT_02', 'CAT_03',
               'CAT_04', 'CAT_05', 'CAT_06']
    return folder_dir, folders

    # preprocess an individual image
    def preprocess_image(img):
    img = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY) #convert to gray
    img = cv2.resize(img, (INPUT_SHAPE, INPUT_SHAPE)) #resize
    img = img / 255.0 #normalize pixel values
    return img

    # preprocess an individual set of keypoints
    def preprocess_keypoints(kp, img):
    kp = kp.reshape(-1, 2) #reshape the array
    kp = kp * [INPUT_SHAPE / img.shape[1], INPUT_SHAPE / img.shape[0]] #resize
    kp = kp / 255.0 # normalize values
    return kp

    # load the images and their corresponding key points
    def load_data():
    # set up
    folder_dir, folders = setup_paths()
    images = []
    keypoints = []
    print("Loading data…")
    
    # iterate over the directories
    for folder in folders:
        path = os.path.join(folder_dir, folder, '*.jpg')
        images_paths = sorted(glob.glob(path))
        
        for image_path in images_paths:
            # load image
            img = cv2.imread(image_path)
            
            # load keypoints
            with open(image_path+'.cat', 'r') as f:
                keypoints_text = f.read().strip()
            kp = np.array(keypoints_text.split(' ')[1:], dtype=np.float32)
            
            # Add the image and keypoints to the lists
            keypoints.append(preprocess_keypoints(kp, img))
            images.append(preprocess_image(img))
            
    print("Done.\n")    
    return images, keypoints

    # prepare the data and split the sets
    def prepare_data():
    # load data
    images, keypoints = load_data()
    
    # convert lists into np arrays
    images = np.array(images)
    keypoints = np.array(keypoints)
    
    # reshape the arrays
    images = np.expand_dims(images, axis=-1)
    keypoints = keypoints.reshape(len(keypoints),18)
    
    print("Splitting data into sets…")   
    
    # split the dataset into train and test sets
    train_val_images, test_images, train_val_keypoints, test_keypoints = 
    model_selection.train_test_split(
        images, keypoints, test_size=0.2, random_state=42)

    # split the train and validation sets
    train_images, val_images, train_keypoints, val_keypoints = model_selection.train_test_split(
        train_val_images, train_val_keypoints, test_size=0.2, random_state=42)
    
    print("Done.\n")   
    return train_images, train_keypoints, val_images, val_keypoints, test_images, test_keypoints


    ### BUILING & EVALUATING THE MODEL
    # define the CNN architecture - might improve later
    def create_model():
    print("Creating the model...") 
    model = tf.keras.Sequential([
        # convolution layers
        tf.keras.layers.Conv2D(32, (3, 3), activation=tf.keras.layers.LeakyReLU(alpha=0.1),   
        input_shape=(INPUT_SHAPE, INPUT_SHAPE, 1)),
        tf.keras.layers.MaxPooling2D((2, 2)),
        
        tf.keras.layers.Conv2D(64, (3, 3), activation=tf.keras.layers.LeakyReLU(alpha=0.1)),
        tf.keras.layers.MaxPooling2D((2, 2)),
        
        tf.keras.layers.Conv2D(128, (3, 3), activation=tf.keras.layers.LeakyReLU(alpha=0.1)),
        tf.keras.layers.MaxPooling2D((2, 2)),
        
        tf.keras.layers.Conv2D(256, (3, 3), activation=tf.keras.layers.LeakyReLU(alpha=0.1)),
        tf.keras.layers.MaxPooling2D((2, 2)),
        
        tf.keras.layers.Conv2D(512, (3, 3), activation=tf.keras.layers.LeakyReLU(alpha=0.1)),
        tf.keras.layers.MaxPooling2D((2, 2)),
        
        tf.keras.layers.Flatten(),
        
        # dense layers
        tf.keras.layers.Dense(1024, activation=tf.keras.layers.LeakyReLU(alpha=0.1)),
        tf.keras.layers.Dense(512, activation=tf.keras.layers.LeakyReLU(alpha=0.1)),
        tf.keras.layers.Dense(256, activation=tf.keras.layers.LeakyReLU(alpha=0.1)),
        tf.keras.layers.Dense(128, activation=tf.keras.layers.LeakyReLU(alpha=0.1)),
        tf.keras.layers.Dense(64, activation=tf.keras.layers.LeakyReLU(alpha=0.1)),
        tf.keras.layers.Dense(32, activation=tf.keras.layers.LeakyReLU(alpha=0.1)),
    
        tf.keras.layers.Dense(18)
    ])
    print("Done.\n")   
    return model

    # train the model
    def train_model(train_images, train_keypoints, val_images, val_keypoints):
    #create the model
    model = create_model()
    
    # compile the model
    print("Compiling the model...")
    model.compile(optimizer='adam', loss="mean_squared_error", 
                  metrics=['mae', RootMeanSquaredError(name='rmse'), r2, med, pck])
    print("Done.\n")
    
    # train the model
    print("Training the model...")
    history = model.fit(train_images, train_keypoints, 
                        epochs=EPOCHS, batch_size=BATCH_SIZE,
                        validation_data=(val_images, val_keypoints))
    print("Done.\n")
    
    # save the model
    **path = 'models-kp/model_' + ITERATION**
    model.save(path)
    return history, model

Save model path in keypoints.py. No matter what number I change the iteration number to, this time I get the error An error occurred: No file or directory found at models-kp/model_(NO).Can anyone help? This is important to me. Thanks in advance.

I want to analyze my pics with this trained model. I edit the code but its not working.