U yh#@sddlmZmZddlmZddlZddlmmmZddl m Z ddl m Z ddlm Z mZddlmZddlmZdgZe je eeeeeffe jd d dZdS) )DictUnionfnmatchN) DeviceMesh)is_rng_supported_meshTensorParallelRNGTracker)_validate_tp_mesh_dim) ParallelStyleparallelize_module)module device_meshparallelize_planreturnc stjdt|t|rLttjtsLt|j t_tjj |dddtj_ t|t rb| ||St|tr|D]\}}|d}t|dkrtd|rv|dtfdd |}|D]4\}}|rd|} t||| |iqt|||qqqv|Std t|d d S) a Apply Tensor Parallelism in PyTorch by parallelizing modules or sub-modules based on a user-specified plan. We parallelize module or sub_modules based on a parallelize_plan. The parallelize_plan contains :class:`ParallelStyle`, which indicates how user wants the module or sub_module to be parallelized. User can also specify different parallel style per module fully qualified name (FQN). Note that ``parallelize_module`` only accepts a 1-D :class:`DeviceMesh`, if you have a 2-D or N-D :class:`DeviceMesh`, slice the DeviceMesh to a 1-D sub DeviceMesh first then pass to this API(i.e. ``device_mesh["tp"]``) Args: module (:class:`nn.Module`): Module to be parallelized. device_mesh (:class:`DeviceMesh`): Object which describes the mesh topology of devices for the DTensor. parallelize_plan (Union[:class:`ParallelStyle`, Dict[str, :class:`ParallelStyle`]]): The plan used to parallelize the module. It can be either a :class:`ParallelStyle` object which contains how we prepare input/output for Tensor Parallelism or it can be a dict of module FQN and its corresponding :class:`ParallelStyle` object. Return: A :class:`nn.Module` object parallelized. Example:: >>> # xdoctest: +SKIP("distributed") >>> from torch.distributed.tensor.parallel import parallelize_module, ColwiseParallel >>> from torch.distributed.device_mesh import init_device_mesh >>> >>> # Define the module. >>> m = Model(...) >>> tp_mesh = init_device_mesh("cuda", (8,)) >>> m = parallelize_module(m, tp_mesh, {"w1": ColwiseParallel(), "w2": RowwiseParallel()}) >>> .. note:: For complex module architecture like Attention, MLP layers, we recommend composing different ParallelStyles together (i.e. ``ColwiseParallel`` and ``RowwiseParallel``) and pass as a parallelize_plan, to achieves the desired sharding computation. z4torch.distributed.tensor.parallel.parallelize_modulei)Z base_seedF.rz9Expect module path to be non-empty, but got empty string!cst|dS)Nrr)tZatomW/var/www/html/venv/lib/python3.8/site-packages/torch/distributed/tensor/parallel/api.pyez$parallelize_module..zLExpect Union[ParallelStyle, Dict[str, ParallelStyle]] for parallelize_plan, z found!N)torchZ_CZ_log_api_usage_oncer r isinstancerandomZ _rng_trackerrZ device_typeZ _manual_seedZdistribute_region_enabledr _applydictitemssplitlen ValueErrorpopfilterZnamed_childrenjoinr TypeErrortype) r r r module_pathZparallelize_styleZ path_splitsZmatched_children_ submoduleZ leaf_pathrrrr sB.            )typingrrrrZ torch.distributed._tensor.random distributedZ_tensorrZtorch.nnnnZtorch.distributed._tensorrrrZ(torch.distributed.tensor.parallel._utilsr Z'torch.distributed.tensor.parallel.styler __all__Modulestrr rrrrs