RRWM

class src.qap_solvers.rrwm.RRWM(max_iter=50, sk_iter=20, alpha=0.2, beta=30)

RRWM solver for graph matching (QAP), implemented by power iteration with Sinkhorn reweighted jumps. Parameter: maximum iteration max_iter Input: input matrix M

maximum size of source graph num_src sizes of source graph in batch ns_src sizes of target graph in batch ns_tgt (optional) initialization vector v0. If not specified, v0 will be initialized with all 1.

Output: computed eigenvector v

forward(M, num_src, ns_src, ns_tgt, v0=None)

training: bool