Array Broadcast error after few batch training

[draxler1]

Hi, I am reproducing your training process. However, at first epoch, after several interaction following error is popping up. Help me.

Epoch 1/50
25/75 [=========>....................] - ETA: 5:53 - loss: 0.5275 - accuracy: 0.7709
---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-4-9a7e0875bfca> in <module>()
    116             epochs=50,
    117             initial_epoch=0,
--> 118             callbacks=[checkpoint_period, reduce_lr])
    119 
    120     model.save_weights(log_dir+'last1.h5')

9 frames
<ipython-input-4-9a7e0875bfca> in generate_arrays_from_file(lines, batch_size)
     46             # Restart after reading a cycle
     47             i = (i+1) % n
---> 48         yield (np.array(X_train),np.array(Y_train))
     49 
     50 def  loss ( y_true , y_pred ):

ValueError: could not broadcast input array from shape (416,416,3) into shape (416,416)