Setting up a data generator for learning a hidden layer

Question

Long time lurker, first-time poster here :)

I am trying to build a CNN in python that will take as input an image and output 25 continuous values predicted from said image. I have a directory containing subdirectories each with the image as a .jpg and the corresponding 25 values stored as a .csv. I have done something similar in the past by reading my images and labels in as numpy arrays but thought it wise to try to learn how to do this with data generators. My code so far is below but as far as I can tell it is not loading the label sets since my model keeps returning an error indicating there is no gradient for it to train on.

I am doing everything in TensorFlow, but am not necessarily married to it. Likewise, I can regenerate how my images and labels are stored so if the way I did it is the problem I am happy to change that structure as well.

Any help or suggestions is appreciated!

Using plt.imshow(train_generator[0][0]) I can confirm the images are loaded correctly.

The specific error I am returning ends with

ValueError: No gradients provided for any variable:

The relevant code I have so far is:

import os, shutil, csv
import tensorflow as tf
from tensorflow.keras.preprocessing.image import ImageDataGenerator

def data_generator(data_dir, batch_size):

# Set up the data generator

datagen = ImageDataGenerator(rescale=1./255)

    # Read in the data from the specified directory
    generator = datagen.flow_from_directory(
        data_dir,
        target_size=image_size,
        batch_size=batch_size,
        class_mode=None,  # since the labels are stored in separate .csv files
        shuffle=True
    )
    
    # Iterate over the data
    for data_batch, _ in generator:
        # Read in the corresponding .csv file for each image
        labels_batch = []
        for i, filename in enumerate(data_batch):
            # Get the file name without the extension
            file_base = os.path.splitext(filename)[0]
    
            # Read in the labels from the .csv file
            with open(file_base + '.csv', 'r') as csv_file:
                reader = csv.reader(csv_file)
                labels_batch.append(next(reader))
        yield data_batch, labels_batch

train_generator = data_generator(os.path.join(data_dir, 'train'), batch_size)

Skipping a little

# Compile the model

model.compile(optimizer='adam', loss='mean_squared_error', metrics=\['accuracy'\])

# Train the model using the data generators

history = model.fit_generator(
train_generator,
steps_per_epoch=len(train_generator),
epochs=10,
validation_data=val_generator,
validation_steps=len(val_generator)
)

# Evaluate the model on the testing data

test_loss, test_acc = model.evaluate_generator(test_generator, steps=len(test_generator))