From ab70cdc39c18fba9c458c93e19fba84a10bdb74d Mon Sep 17 00:00:00 2001 From: Jaroslav Fowkes Date: Wed, 11 Sep 2024 11:43:18 +0200 Subject: [PATCH] LSQML: first replicate ML --- ptypy/custom/LSQML.py | 136 ++++++++++++++++++++++-------------------- 1 file changed, 70 insertions(+), 66 deletions(-) diff --git a/ptypy/custom/LSQML.py b/ptypy/custom/LSQML.py index d2247f77..13cbce52 100644 --- a/ptypy/custom/LSQML.py +++ b/ptypy/custom/LSQML.py @@ -83,6 +83,7 @@ def __init__(self, ptycho_parent, pars=None): self.pr_tmin = None # Other + self.tmin = None # for ML step only self.ML_model = None self.ptycho.citations.add_article( @@ -118,6 +119,9 @@ def engine_initialize(self): self.pr_buf = self.pr.copy(self.pr.ID + '_buf', fill=0.) self.pr_tmin = {} # need scalar per named pod + # for ML step only + self.tmin = 1. + self._initialize_model() def _initialize_model(self): @@ -176,11 +180,17 @@ def engine_iterate(self, num=1): self.ob_grad << new_ob_grad self.pr_grad << new_pr_grad - # Next conjugate - self.ob_h *= bt - self.ob_h -= self.ob_grad - self.pr_h *= bt + # # Next conjugate + # self.ob_h *= bt + # self.ob_h -= self.ob_grad + # self.pr_h *= bt + # self.ob_h -= self.ob_grad + + # Next ML conjugate + self.ob_h *= bt / self.tmin self.ob_h -= self.ob_grad + self.pr_h *= bt / self.tmin + self.pr_h -= self.pr_grad # ############################ # # Compute steepest descent # @@ -190,20 +200,20 @@ def engine_iterate(self, num=1): # self.pr_h << new_pr_grad # self.pr_h *= -1 - ########################## - # Average direction (25) # - ########################## - self.ob_nrm += 1e-6 - self.ob_h /= self.ob_nrm - self.pr_nrm += 1e-6 - self.pr_h /= self.pr_nrm + # ########################## + # # Average direction (25) # + # ########################## + # self.ob_nrm += 1e-6 + # self.ob_h /= self.ob_nrm + # self.pr_nrm += 1e-6 + # self.pr_h /= self.pr_nrm - ######################## - # Compute step lengths # - ######################## - t2 = time.time() - self.ML_model.compute_step_lengths() - ts += time.time() - t2 + # ######################## + # # Compute step lengths # + # ######################## + # t2 = time.time() + # self.ML_model.compute_step_lengths() + # ts += time.time() - t2 # ################################ # # Take weighted mean step (27) # @@ -235,9 +245,9 @@ def engine_iterate(self, num=1): ################ t3 = time.time() B = self.ML_model.poly_line_coeffs() - tmin = self.ptycho.FType(-.5 * B[1] / B[2]) - self.ob_h *= tmin - self.pr_h *= tmin + self.tmin = self.ptycho.FType(-.5 * B[1] / B[2]) + self.ob_h *= self.tmin + self.pr_h *= self.tmin self.ob += self.ob_h self.pr += self.pr_h tu += time.time() - t3 @@ -347,9 +357,45 @@ def new_grad(self): def compute_step_lengths(self): """ - Compute optimization step lengths according to the noise model. + Compute LSQML optimization step lengths according. """ - raise NotImplementedError + + # Outer loop: through diffraction patterns + for dname, diff_view in self.di.views.items(): + if not diff_view.active: + continue + + # Third pod loop: calculate real-space step lengths + for name, pod in diff_view.pods.items(): + if not pod.active: + continue + + # Get xi + xi = pod.exit + + # Get update directions + ob_h = self.ob_h[pod.ob_view] + pr_h = self.pr_h[pod.pr_view] + + # Compute cross-terms + ob_h_pr = ob_h * pod.probe + pr_h_ob = pr_h * pod.object + + # Calculate real-space step lengths (22) + M = np.zeros((2,2), dtype=np.cdouble) + rhs = np.zeros(2, dtype=np.double) + M[0,0] = np.sum(u.abs2(ob_h_pr)) + 1e-6 + M[1,1] = np.sum(u.abs2(pr_h_ob)) + 1e-6 + M[0,1] = np.sum(ob_h_pr * pr_h_ob.conj()) + M[1,0] = np.sum(ob_h_pr.conj() * pr_h_ob) + rhs[0] = np.sum(np.real(xi * ob_h_pr.conj())) + rhs[1] = np.sum(np.real(xi * pr_h_ob.conj())) + #self.ob_tmin[name], self.pr_tmin[name] = np.linalg.solve(M, rhs) + self.ob_tmin[name], self.pr_tmin[name] = sp.linalg.solve(M, rhs, assume_a='her') + + # # Calculate approx real-space step lengths (23) + # self.ob_tmin[name] = np.sum(np.real(xi * ob_h_pr.conj())) / (np.sum(u.abs2(ob_h_pr)) + 1e-6) + # self.pr_tmin[name] = np.sum(np.real(xi * pr_h_ob.conj())) / (np.sum(u.abs2(pr_h_ob)) + 1e-6) def new_step(self): """ @@ -421,8 +467,8 @@ def __init__(self, MLengine): for name, di_view in self.di.views.items(): if not di_view.active: continue - self.weights[di_view] = (self.Irenorm * di_view.pod.ma_view.data - / (1./self.Irenorm + di_view.data)) + self.weights[di_view] = (di_view.pod.ma_view.data + / (1 + di_view.data)) def __del__(self): """ @@ -708,48 +754,6 @@ def new_grad(self): return error_dct - def compute_step_lengths(self): - """ - Compute optimization step lengths according to a Euclidean noise model. - """ - - # Outer loop: through diffraction patterns - for dname, diff_view in self.di.views.items(): - if not diff_view.active: - continue - - # Third pod loop: calculate real-space step lengths - for name, pod in diff_view.pods.items(): - if not pod.active: - continue - - # Get xi - xi = pod.exit - - # Get update directions - ob_h = self.ob_h[pod.ob_view] - pr_h = self.pr_h[pod.pr_view] - - # Compute cross-terms - ob_h_pr = ob_h * pod.probe - pr_h_ob = pr_h * pod.object - - # Calculate real-space step lengths (22) - M = np.zeros((2,2), dtype=np.cdouble) - rhs = np.zeros(2, dtype=np.double) - M[0,0] = np.sum(u.abs2(ob_h_pr)) + 1e-6 - M[1,1] = np.sum(u.abs2(pr_h_ob)) + 1e-6 - M[0,1] = np.sum(ob_h_pr * pr_h_ob.conj()) - M[1,0] = np.sum(ob_h_pr.conj() * pr_h_ob) - rhs[0] = np.sum(np.real(xi * ob_h_pr.conj())) - rhs[1] = np.sum(np.real(xi * pr_h_ob.conj())) - #self.ob_tmin[name], self.pr_tmin[name] = np.linalg.solve(M, rhs) - self.ob_tmin[name], self.pr_tmin[name] = sp.linalg.solve(M, rhs, assume_a='her') - - # # Calculate approx real-space step lengths (23) - # self.ob_tmin[name] = np.sum(np.real(xi * ob_h_pr.conj())) / (np.sum(u.abs2(ob_h_pr)) + 1e-6) - # self.pr_tmin[name] = np.sum(np.real(xi * pr_h_ob.conj())) / (np.sum(u.abs2(pr_h_ob)) + 1e-6) - def poly_line_coeffs(self): """ Compute the coefficients of the polynomial for line minimization