raise RuntimeError('You must compile your model before using it.') in Keras at fit.generator

Question

I am working on an image caption model. While training the model I am getting a compile error. I tried many solutions but can't seem to solve the problem. Also, my code was in old Keras version. I am using Keras 2.2.4: Please see below the code that in written in old Keras version:

import numpy as np
import pandas as pd
from keras.models import Sequential, Model
from keras.layers import LSTM, Embedding, TimeDistributed, Dense, RepeatVector, Activation
from keras.layers.merge import Concatenate
from keras.preprocessing import image, sequence
import pickle as pickle

def create_model(self, ret_model = False):  
        image_model = Sequential()
        image_model.add(Dense(EMBEDDING_DIM, input_dim = 4096, activation='relu'))
        image_model.add(RepeatVector(self.max_length))

        lang_model = Sequential()
        lang_model.add(Embedding(self.vocab_size, 256, input_length=self.max_length))
        lang_model.add(LSTM(256,return_sequences=True))
        lang_model.add(TimeDistributed(Dense(EMBEDDING_DIM)))

        model = Sequential()
        model.add(Merge([image_model, lang_model], mode='concat'))
        model.add(LSTM(1000,return_sequences=False))
        model.add(Dense(self.vocab_size))
        model.add(Activation('softmax'))

        print ("Model created!")

        if(ret_model==True):
            return model

        model.compile(loss='categorical_crossentropy', optimizer='rmsprop', metrics=['accuracy'])
        return model

I changed the code to like this. The libraries are the same as i showed in the first code. Please see below the changed code:

def create_model(self, ret_model = False):
        image_model = Sequential()
        image_model.add(Dense(EMBEDDING_DIM, input_dim = 4096, activation='relu'))
        image_model.add(RepeatVector(self.max_length))

        lang_model = Sequential()
        lang_model.add(Embedding(self.vocab_size, 256, input_length=self.max_length))
        lang_model.add(LSTM(256, return_sequences=True))
        lang_model.add(TimeDistributed(Dense(EMBEDDING_DIM)))

        model = Sequential()
        #model.add(Merge([image_model, lang_model], mode='concat'))
        model.add(Concatenate([image_model, lang_model])) #changed the above commented line to this
        model.add(LSTM(1000,  return_sequences=False))
        model.add(Dense(self.vocab_size))
        model.add(Activation('softmax'))

        print ("Model created!")

        if(ret_model==True):
            return model

        model.compile(loss='categorical_crossentropy', optimizer='rmsprop', metrics=['accuracy'])
        return model

When I use this above-changed code and train my model I get a runtime error which says I must compile my code. Though I am compiling my code. Please see the error below:

Using TensorFlow backend.
413439
WARNING:tensorflow:From C:\Users\UserName\Anaconda3\lib\site- 
packages\tensorflow\python\framework\op_def_library.py:263: colocate_with (from 
tensorflow.python.framework.ops) is deprecated and will be removed in a future version.
Instructions for updating:
Colocations handled automatically by placer.
Model created!
Traceback (most recent call last):
  File "D:\Image-caption\Image-Captioning-master\train.py", line 14, in <module>
    train(int(sys.argv[1]))
  File "D:\Image-caption\Image-Captioning-master\train.py", line 9, in train
    model.fit_generator(sd.data_process(batch_size=batch_size), 
 steps_per_epoch=sd.no_samples/batch_size, epochs=epoch, verbose=2, callbacks=None)
  File "C:\Users\UserName\Anaconda3\lib\site-packages\keras\legacy\interfaces.py", line 91, in 
wrapper
    return func(*args, **kwargs)
  File "C:\Users\UserName\Anaconda3\lib\site-packages\keras\engine\training.py", line 1418, in 
  fit_generator
  initial_epoch=initial_epoch)
  File "C:\Users\UserName\Anaconda3\lib\site-packages\keras\engine\training_generator.py", line 40, 
  in fit_generator
    model._make_train_function()
  File "C:\Users\UserName\Anaconda3\lib\site-packages\keras\engine\training.py", line 496, in 
  _make_train_function
    raise RuntimeError('You must compile your model before using it.')
RuntimeError: You must compile your model before using it.

This is my train.py file code:

import SceneDesc
import sys

def train(epoch):
    sd = SceneDesc.scenedesc()
    model = sd.create_model()
    batch_size = 512
    model.fit_generator(sd.data_process(batch_size=batch_size), 
    steps_per_epoch=sd.no_samples/batch_size, epochs=epoch, verbose=2, callbacks=None)
    model.save('Output/Model.h5', overwrite=True)
    model.save_weights('Output/Weights.h5',overwrite=True)

if __name__=="__main__":
    train(int(sys.argv[1]))

I am a beginner and in a learning phase. I am trying to solve this probe for quite some time but can't seem to solve the issue. I have also provide the link of my files in dpaste. Here is the link: https://dpaste.de/t2yU and https://dpaste.de/cVGP ... Though these are only two files. I tried my best to ask a clear question. I hope to get a positive response. :) Any help will be apreciated.