Closed alexblnn closed 2 years ago
Since PR #22 we use a loop in the calibration function to apply the inverse templates :
tkInv_all = np.array([t_inv(p0[:, i], i + 1, p) for i in range(p)]).T
Perhaps this could be avoided. We can see that easily in the linear case, but the general case is not trivial (for example, for beta templates)
As far as I can see, the fix would be to have t_inv accept an array as a second argument. t_inv would perform the loop internally; I guess that for some templates it can be vectorized easily.
t_inv
too low priority.
Since PR #22 we use a loop in the calibration function to apply the inverse templates :
tkInv_all = np.array([t_inv(p0[:, i], i + 1, p) for i in range(p)]).T
Perhaps this could be avoided. We can see that easily in the linear case, but the general case is not trivial (for example, for beta templates)