cotengra.experimental.path_compressed_mcts#
Compressed contraction tree search using monte carlo tree search.
Module Contents#
Classes#
- class cotengra.experimental.path_compressed_mcts.Node(hg, nid_path, size, local_score, forward_score)#
- __slots__ = ('hg', 'n', 'graph_key', 'nid_path', 'size', 'local_score', 'forward_score', 'mean', 'count',...#
- update(x)#
Report the score
x, presumably from a child node, updating this nodes score.
- __hash__()#
Return hash(self).
- __lt__(other)#
Return self<value.
- __repr__()#
Return repr(self).
- class cotengra.experimental.path_compressed_mcts.MCTS(chi, T=0.1, prune=True, optimize=None, optimize_factory=False, seed=None)#
- property ssa_path#
- property path#
- __repr__()#
Return repr(self).
- setup(inputs, output, size_dict)#
- get_ssa_path()#
Convert unique node identifiers to ssa.
- check_node(node)#
- delete_node(node)#
- backprop(node)#
- simulate_node(node)#
- simulate_optimized(node)#
- is_deadend(node)#
- descend()#
- run(inputs, output, size_dict)#
- search(inputs, output, size_dict)#
- __call__(inputs, output, size_dict)#