Overflow in python

#y=x^4-x^3
def function(x):
    return x**4-2*(x**3)+1

# minimum: y=-11/16, x=3/2
defderiv(x):
    return4*(x**3) - 6*(x**2)


def GD(deriv, init, eta,iter=100):
    '''
    —param deriv:derivative of the function you want to optimize
    —param init:start point, initial value
    —parameta:leaning rate, step size
    —return:history
    '''
    eps = 1e-5
    x = init
    x_history = [init]
    for i in range (iter):
        x_=x-eta*deriv(x)
        ifabs(x-x_)<eps:#convergence condition
            break
        x = x_
        x_history.append(x)
    return np.array(x_history)


etas = [0.1, 0.2, 0.4, 0.5]
'''
We check the behavior of Gradient Descent combined with the different eta
'''

for i, etain enumerate (etas):
    tracks = GD (deriv, 2.0, eta)

    plt.subplot(2,2,(i+1))

    plt.title("eta:"+str(eta)+", iter:"+str(len(tracks)))))

    x = np.range (-0.3, 1.0, 0.01)
    y = function(x)
    plt.plot(x,y,linestyle="-",c="black")
    plt.plot(tracks, function(tracks), 'x', c="r")

#plt.tight_layout()
plt.show()