loss_fns

#Training replacement

d = 0 #from algorithm 2, used in computing loss
#alpha = .33 #learning rate

#def nt_grad(curr_d):
    #return -1*alpha*(curr_d/BATCH_SIZE)

def compute_loss(model, x, y, training):
    out = model(x, training=training)
    loss = loss_object(y_true=y, y_pred=out)
    
    #with tf.GradientTape() as tape:
        
    
    return loss


def get_grad(model, x, y):
    with tf.GradientTape() as tape:
        loss = compute_loss(model, x, y, training=True)
    return loss, tape.gradient(loss, model.trainable_variables)

def model_2_compute_loss(model, x, y, loss_value):#, training, sl, L1, L2, L3):
    print("model_2: print x shape in compute loss f'n" + str(np.shape(x)))
    print("model_2: print y shape in compute loss f'n" + str(np.shape(y)))
    print("model_2: print loss_value shape in compute loss f'n" + str(np.shape(loss_value)))
    out = model(x, training=True)
    loss = loss_value #model_2_loss_function(sl, L1, L2, L3)
    #return tf.convert_to_tensor(loss) <-What Jonathan saw
    return loss

def model_2_get_grad(model, x, y, loss_value_tensor): #, sl, L1, L2, L3):
    #count = 0
    print("model_2: print loss_value_tensor in get grad f'n" + str(np.shape(loss_value_tensor)))
    with tf.GradientTape(persistent=True) as tape:
        
        for count in range(BATCH_SIZE):
        
            print( "loss value tensor at" + " " + str(count) + ": " + str(loss_value_tensor[count]))
            loss = model_2_compute_loss(model, x, y, loss_value_tensor[count]) #, True, sl, L1, L2, L3) #training = True
            #count = count + 1
            #returning only the first argument??
            #print(repr(model.trainable_variables))
            #print(repr(loss))
            #tape.gradient(loss, model.trainable_variables) <- I think this was for printouts
    return loss, tape.gradient(loss, model.trainable_variables)