Fast way to implement a polynomial regression on each pandas dataframe row

I have the following pandas dataframe:

df = pd.DataFrame({0: [11, 12, 31], 1: [6, 14, 27], 2: [11, 24, 21], 3: [1, 24, 20]})

    0   1   2   3
0  11   6  11   1
1  12  14  24  24
2  31  27  21  20

For each row at the time, I want to implement a polynomial regression, where column names are X and row values are Y.

I know I can use iterrows:

x=(df.columns).to_numpy()

for index, row in df.iterrows():
    print(np.polyfit(x,row,2))

which produces:

[-1.25  1.25  9.75]
[-0.5  6.1 11.1]
[ 0.75 -6.15 31.35]

but this can take a long time on large dataframes.
Is there a faster way to do this? Thanks

>Solution :

polyfit can take a 2d y-param of the same length with x, so:

np.polyfit(df.columns, df.T, 2).T

gives:

array([[-1.25,  1.25,  9.75],
       [-0.5 ,  6.1 , 11.1 ],
       [ 0.75, -6.15, 31.35]])