In Seaborn's FacetGrid how to adjust the scale of the x axes to the range of available data

Question

I have the following Seaborn FaceGrid output:

generated with a snippet like this:

imax = 3

df = pd.DataFrame(columns = ['x','y','i'])
for i in range(imax):
    # Creating vectors X and Y
    x = np.linspace(0, 1/(2 ** i),101)
    y = x ** x
    df = df.append(pd.DataFrame({'x': x, 'y': y, 'i':i}), ignore_index = True)

g = sns.FacetGrid(df, col='i')
g = g.map(sns.lineplot, 'x', 'y')

import matplotlib.pyplot as plt
plt.show()

(See this gist for the full source.)

In the grid I would like the x axes always be adjusted to the available x values for the respective "i". How can I do this?

What I have tried

I called set_xbound on the axes objects:

for i,b in enumerate(g.axes[0]):
    b.set_xbound(0, 1/(2 ** i))

but this sets the x-scale to [0,0.25] for all subplots, while I obviously want for the first subplot [0,1], then [0.5] and only for the last subplot [0,0.25].