Writing a test_model function in pytorch

hi guys i wanna write a function test_model(model, criterion):
It should print the loss and the test set accuracy , and should return y_true and y_pred as NumPy arrays.

so i tried the following :

def test_model(model_ft, test_dataloaders,criterion,device):

# set model to evaluate model
# Iterate over data.
for inputs, labels in dataloaders[phase]:
    inputs = inputs.to(device)
    labels = labels.to(device)

and now my question how do i get y_pred and y_true and the rest things?
We are in the framework of transfer learning in pytorch so the function i wanna write is the same as the train_model see(https://pytorch.org/tutorials/beginner/transfer_learning_tutorial.html) but somewhat edited for the prediction purpose. I also thought of concatenate the y_pred and y_true after each batch but i really cant put this into reality could you plz help me out ? Thanks