Deeploy.DeeployTypes.NetworkOptimizer

class Deeploy.DeeployTypes.NetworkOptimizer(passes: List[TopologyOptimizationPass], name: str = 'TopologyOptimizer')

Bases: TopologyOptimizer

Wrapper class to run multiple NetworkOptimizationPasses sequentially

Methods

__init__(passes: List[TopologyOptimizationPass], name: str = 'TopologyOptimizer')

__init__(passes[, name])
optimize(ctxt, graph)	Apply passes sequentially

optimize(ctxt: NetworkContext, graph: Graph) → Tuple[NetworkContext, Graph]

Apply passes sequentially

Parameters:

• ctxt (NetworkContext) – Current NetworkContext

• graph (gs.Graph) – Current Neural Network graph

Returns:

Tuple[NetworkContext, gs.Graph] – Update context and graph

Return type:

# type: ignor