Searched refs:y_pos_flat (Results 1 – 1 of 1) sorted by relevance
247 auto y_pos_flat = y_pos[y_idx].flat<Y_T>(); in ComputeNumericJacobianTranspose() local255 (y_pos_flat(c) - y_neg_flat(c)) / scale, in ComputeNumericJacobianTranspose()