hG%SSKrSSKJrJrJr SSKrSSKJr 0r\\\\\RR\/\RR44\ S'Sr g)N)CallableDictType) AOBaseConfig_QUANTIZE_CONFIG_HANDLERcF^[R"T5U4Sj5nU$)a" A decorator to register a transform function to map from a workflow configuration (child of `AOBaseConfig`) to a function that transforms a `torch.nn.Module` according to the specified configuration. For example:: # user facing code class WorkflowFooConfig(AOBaseConfig): ... # configuration for workflow `Foo` is defined here bar = 'baz' # non user facing code @register_quantize_module_handler(WorkflowFooConfig) def _transform( mod: torch.nn.Module, config: WorkflowFooConfig, ) -> torch.nn.Module: # the transform is implemented here, usually a tensor sublass # weight swap or a module swap ... # then, the user calls `quantize_` with a config, and `_transform` is called # under the hood by `quantize_. c>U[T'U$)N)r)func config_types _/home/james-whalen/.local/lib/python3.13/site-packages/torchao/quantization/transform_module.py decorator3register_quantize_module_handler..decorator/s04 - ) functoolswraps)r r s` r register_quantize_module_handlerrs(8__[!" r) rtypingrrrtorchtorchao.core.configrrnnModule__annotations__rrr rsa '' , $ ehhoo| ,ehhoo =>@ !r