Question about the data entered in the testing phase

[TMRajeh]

I have a question about the data entered in the testing process for the model. How was the data entered during the test phase and how was it passed to the model?

X[j, :, :] = self.X[range(
batch_idx[j] + self.train_timesteps - self.T, batch_idx[j] + self.train_timesteps - 1), :]
y_history[j, :] = self.y[range(
batch_idx[j] + self.train_timesteps - self.T, batch_idx[j] + self.train_timesteps - 1)]

X here will enter to the encoder and y_history will enter to the decoder.
I'm confused here because I'm new in this field.
how you can pass X and y_history to the model? in practice they must be unknown and you need to predict them? you only have the last value predicted during the training process.
I think here we should take the predicted value from the previous step and add them to the data and use them as input to the next step I don't know if I'm alright or not

Can you explain to me how the test process was done for the model?