Input 0 of layer "sequential" is incompatible with the layer

I try to test a neural network for a game but always when i run the script i get this error:

ValueError: Input 0 of layer "sequential" is incompatible with the layer: expected shape=(None, 24), found shape=(None, 8410)

The model takes data from a game in order to predict the correct key to press.I realy don't understand what i m doing wrong

Here is the code for the neural network program:

import pickleimport tensorflow as tfimport matplotlib.pyplot as pltimport numpy as npimport kerasimport keras.metricsimport keras.backend as Kimport keras.layersdefault_sequence = [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1]def f1(y_true, y_pred):    def recall_function(y_true, y_pred):        true_positives = K.sum(K.round(K.clip(y_true * y_pred, 0, 1)))        possible_positives = K.sum(K.round(K.clip(y_true, 0, 1)))        recall_value = true_positives / (possible_positives + K.epsilon())        return recall_value    def precision_function(y_true, y_pred):        true_positives = K.sum(K.round(K.clip(y_true * y_pred, 0, 1)))        predicted_positives = K.sum(K.round(K.clip(y_pred, 0, 1)))        precision_value = true_positives / (predicted_positives + K.epsilon())        return precision_value    precision = precision_function(y_true, y_pred)    recall = recall_function(y_true, y_pred)    return 2 * ((precision * recall) / (precision + recall + K.epsilon()))inputPath = "Dataset/stateData0.pickle"outputPath = "Dataset/outputData0.pickle"f = open(inputPath, "rb")g = open(outputPath, "rb")X = []y = []try:    while True:        example = pickle.load(f)        a = example[0]        a += example[2]        X.append(a)        y.append(pickle.load(g))except:    passf.close()g.close()y = list(filter(None, y))while len(X) != len(y):    y.append(default_sequence)print(len(X))print(len(y))for i in range(len(y)):    print(y[i])    print("\n")X = np.asarray(X, dtype=np.float32)y = np.asarray(y, dtype=np.float32)training_size = 3000x_train = X[:training_size]y_train = y[:training_size]x_test = X[training_size:]y_test = y[training_size:]a = 0for i in range(len(y_train)):    if y_train[i][13] == 1:    a += 1print("node imbalance:", a, "/", len(y_train))print("X SHAPE:", X.shape)print("Y SHAPE:", y.shape)print("Train SHAPES")print(x_train.shape)print(y_train.shape)print("Test SHAPES")print(x_test.shape)print(y_test.shape)config = tf.compat.v1.ConfigProto()config.gpu_options.allow_growth = Truesession = tf.compat.v1.Session(config=config)model = keras.models.Sequential()model.add(keras.layers.Dense(128, activation=tf.nn.relu))model.add(keras.layers.Dense(128, activation=tf.nn.relu))model.add(keras.layers.Dense(128, activation=tf.nn.relu))model.add(keras.layers.Dense(14, activation=tf.nn.sigmoid))model.compile(optimizer="adam", loss="categorical_crossentropy", metrics=[f1])class_weights = {            0: 1,            1: 1,            2: 1,            3: 1,            4: 3,            5: 3,            6: 3,            7: 2,            8: 2,            9: 2,            10: 2,            11: 1,            12: 1,            13: 1        }model.fit(x_train, y_train, epochs=10, batch_size=16, class_weight=class_weights, validation_data=[x_test,y_test])val_loss, val_prec = model.evaluate(x_test, y_test)print("Validation loss: ", val_loss, "Validation_accuracy: ", val_prec)model.save("Models/nk_ff_fighter.model")print('Model Saved')print(model.summary())

OUTPUT:

node imbalance: 52 / 3000X SHAPE: (3349, 24)Y SHAPE: (3349, 14)Train SHAPES(3000, 24)(3000, 14)Test SHAPES(349, 24)(349, 14)Model SavedModel: "sequential"_________________________________________________________________Layer (type)                Output Shape              Param #   ================================================================= dense (Dense)               (None, 128)               3200       dense_1 (Dense)             (None, 128)               16512      dense_2 (Dense)             (None, 128)               16512      dense_3 (Dense)             (None, 14)                1806      =================================================================Total params: 38,030Trainable params: 38,030Non-trainable params: 0_________________________________________________________________None

Here is the code for the test program:

import numpy as npfrom PIL import ImageGrabimport cv2import timefrom Scripts.OutputControlScript import switch_guard, UPfrom Scripts.StateDetector import checkCurrentGuardInput, generate_feature_listfrom Scripts.processInputToGame import inputToGameimport keyboardimport tensorflow as tfimport kerasimport keras.backend as KmaxHealth = 6running = Falsegpu_options = tf.compat.v1.GPUOptions(per_process_gpu_memory_fraction=0.333)sess = tf.compat.v1.Session(config=tf.compat.v1.ConfigProto(gpu_options=gpu_options))def toggle_running(key):    global running    if running:        print("Neural Knight OFF.")        running = False    else:        print("Neural Knight ON...")def read_data(state):    if state[3] == 1:        print("The oponent guard is right")    if state[4] == 1:        print("The oponent guard is top")    if state[5] == 1:        print("The oponent guard is left")def f1(y_true, y_pred):    def recall_function(y_true, y_pred):        true_positives = K.sum(K.round(K.clip(y_true * y_pred, 0, 1)))        possible_positives = K.sum(K.round(K.clip(y_true, 0, 1)))        recall_value = true_positives / (possible_positives + K.epsilon())        return recall_value    def precision_function(y_true, y_pred):        true_positives = K.sum(K.round(K.clip(y_true * y_pred, 0, 1)))        predicted_positives = K.sum(K.round(K.clip(y_pred, 0, 1)))        precision_value = true_positives / (predicted_positives + K.epsilon())        return precision_value    precision = precision_function(y_true, y_pred)    recall = recall_function(y_true, y_pred)    return 2 * ((precision * recall) / (precision + recall + K.epsilon()))def main():    global running    keyboard.on_release_key("v", toggle_running)    for i in list(range(1))[::-1]:        print(i + 1)        time.sleep(1)        print("Loading model...")    switch_guard(UP)    currentGuardDirection = UP    model = keras.models.load_model('path_to_model', custom_objects={"f1": f1})    print("Neural Warrior is now Active")    prediction = np.array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1])    prevFrames = 600    prevPrediction = []    for i in range(prevFrames):        prevPrediction.append(prediction)    print(prediction)    while True:        # screen = np.array(ImageGrab.grab(bbox=(0, 130, 1024, 700)))        screen = cv2.imread('test.jpg')        currentGuardDirection = checkCurrentGuardInput(prediction, currentGuardDirection)        state = generate_feature_list(screen, currentGuardDirection)        extractedState = state[0]        prevPrediction = prevPrediction[1:]        prevPrediction.append(prediction)        predHistory = [y for x in prevPrediction for y in x]        extractedState = np.expand_dims(extractedState, axis=0)        extractedState = np.concatenate((extractedState[0], predHistory), axis=0)        extractedState = np.expand_dims(extractedState, axis=0)        prediction = model.predict(extractedState)[0]        b = np.zeros_like(prediction)        biggest = 0        biggest_Index = -1        for i in range(4, 7):            if prediction[i] > 0.2:                if prediction[i] > biggest:                    biggest = prediction[i]                    biggest_Index = i        if biggest_Index > -1:            b[biggest_Index] = 1        for i in range(0, 4):            if prediction[i] > 0.2:                b[i] = 1        for i in range(7, len(b)):            if prediction[i] > 0.2:                b[i] = 1            inputToGame(b)        print(b)      if keyboard.is_pressed("."):          cv2.destroyAllWindows()          breakmain()

OUTPUT

 ValueError: Input 0 of layer "sequential" is incompatible with the layer: expected shape=(None,  24), found shape=(None, 8410)