o "i'@sddlZddlZddlmZddlmZmZmZddlZddl m Z ddlm Z ddl m ZddlmZdd lmZdd lmZ ed d Zed dZeddZeddZeddddZddZeee ddZededeeddZededdd Zed edeed!Zed"edGd#d$d$e Z!Gd%d&d&e Z"ed'ej#j$fd(d)Z%ed'ej#j$fd*d+Z&ed'ej#j$fd,d-Z'ed'ej#j$fd.d/Z(ej)Gd0d1d1Z*d'ej#j$fd2d3Z+Gd4d5d5Z,dS)6N) import_module)AnyListOptional)#min_cut_rematerialization_partition)_guards)config) ts_compile) aot_autograd)register_debug_backendcCs|jSN)forwardgmfake_tensor_inputsr^/home/ubuntu/SoloSpeech/.venv/lib/python3.10/site-packages/torch/_dynamo/backends/debugging.pyeagersrcsfdd}|S)Nc s4z|WSty}ztjjd|d}~ww)Nz7Unexpected exception when running generated GraphModule) Exceptiontorch_dynamoexcTorchDynamoException)argserrrinners zeager_noexcept..innerrrrrrrreager_noexcepts rcs4ddlm}fdd}||dd|}||S)Nr)make_fxcstjj|Sr rfx Interpreterrunrrrr runnable_gm.z'pre_dispatch_eager..runnable_gmT) pre_dispatch)"torch.fx.experimental.proxy_tensorr print_readable)rrr r&pre_dispatch_gmrrrpre_dispatch_eager*s  r,csddlmfdd}|S)Nr)SchemaCheckModecs<tjj|WdS1swYdSr r!r%r-rrrr>s$zeager_debug..inner)#torch._subclasses.schema_check_moder-rrr.r eager_debug7s r0ts)namecCs tj|Sr )rjitscriptrrrr torchscriptEs r5csfdd}d|_|S)Ncstj|Sr )rr"r# boxed_runr%fx_grrr$Lr'zboxed_nop..runT) _boxed_call)r8example_inputsr$rr7r boxed_nopKs r;T) fw_compiler partition_fnkeep_inference_input_mutations aot_eager)r2 compiler_fn)r<r>aot_eager_default_partitionerc CsTtjddtttddtjtddd||WdS1s#wYdS)NT)unlift_effect_tokenscSs tdS)Nztorch._inductor.compile_fx)rselect_decomp_tablerrrrosz,aot_eager_decomp_partition..inductor)compiler)r< bw_compilerdecompositionsr=)functorch_configpatchr r; functoolspartialrrrrraot_eager_decomp_partitionhs $rM)r<aot_tsc@ eZdZdS)ReluCompileErrorN__name__ __module__ __qualname__rrrrrPrPc@rO)TestingOnlyCompileErrorNrQrrrrrVrUrVrcCs"|jjD] }|jtjkrtq|Sr )graphnodestargetrrelurPrr:noderrrrelu_compile_error_TESTING_ONLYs  r]cCs4|jjD]}|jtjkrtj|_d|_q||S)N)FReluRuntimeError)rWrXrYrrZ_assertr recompiler[rrrrelu_runtime_error_TESTING_ONLYs  racCs>|jjD]}|jtjkrtj|_|jddf|_q||S)Nrr )rWrXrYrrZaddrr`r[rrr relu_accuracy_error_TESTING_ONLYs  rccCs8|jjD] }|jdkr nq|S|D]}|jstq|S)N call_function)rWrXopis_leafrV)rr:r\trrr#non_leaf_compile_error_TESTING_ONLYs  rhc@seZdZUdZeejjed<e ed<e ed<ee ed<e ed<dZ e eejj ed<dZe eejed <dZe eed <d d ZdS) ExplainOutputzu This is the output of :func:`torch._dynamo.explain()` There is no reason to create this class directly. graphs graph_countgraph_break_count break_reasonsop_countN ops_per_graph out_guards compile_timesc CsXd|jd}|d|jd7}|d|jd7}|d7}t|jD])\}}|d|dd7}|d |jd7}|d 7}|jD] }|d |d7}q@q"|jdurv|d 7}t|jD]\}}|d |dd7}|D] }|d|d7}qjqZ|jdur|d7}t|jD]\}}|d|dd7}|dt |7}q|j dur|d|j d7}|S)Nz Graph Count:  zGraph Break Count: z Op Count: zBreak Reasons: z Break Reason r z: z Reason: z User Stack: z zOps per Graph: z Ops z z Out Guards: z Guard zCompile Times: ) rkrlrn enumeratermreason user_stackrorpstrrq) selfoutputidx break_reason frame_summaryopsreiguardrrr__str__s4    zExplainOutput.__str__)rRrSrT__doc__rrr" GraphModule__annotations__intrrorNoderprGuardrqrvrrrrrris  ricCsT||dd|jjD}|t|7}|||jjr#||j|||||fS)a This function is a utility which processes a torch.fx.GraphModule and accumulates information about its ops, graph breaks, and other details. It is intended to be used by the ExplainWithBackend class and `torch._dynamo.explain()` to provide details from Dynamo's graph capture. Parameters: gm (torch.fx.GraphModule): The GraphModule to be processed. graphs (list): A list that accumulates all the GraphModules processed. op_count (int): The total count of operations in all GraphModules processed so far. ops_per_graph (list): A list that accumulates the operations of each GraphModule. break_reasons (list): A list that accumulates the reasons for breaks in each GraphModule. Returns: tuple: A tuple containing the processed GraphModule, the updated lists of graphs, operations per graph, and break reasons, and the updated operation count. cSsg|] }|jdkr|jqS)rd)rerY).0r\rrr sz)_explain_graph_detail..)appendrWrXlencompile_subgraph_reason graph_break)rrjrnrormr|rrr_explain_graph_details    rc@s8eZdZdZddZdejjfddZde fdd Z d S) ExplainWithBackenda This class is intended to be used as a backend for `torch.compile`. It is composable with other backends. When used in this way, it accumulates information about graph breaks, ops, and other info and provides a string representation summarizing this information. Attributes: backend (str): The name of the backend to use for optimization. graphs (list): A list of the graphs captured by TorchDynamo. op_count (int): The total number of operations in all optimized graphs. break_reasons (list): A list of graph break reasons with stack traces. Example Usage: def fn(x): x = torch.sigmoid(x) return x torch._dynamo.reset() eb = ExplainWithBackend("inductor") optimized_fn = torch.compile(fn, backend=eb) result = optimized_fn(torch.randn(5)) print(eb.output()) cCs,ddlm}|||_g|_d|_g|_dS)Nr )lookup_backendr)registryrbackendrjrnrm)rwrrrrr__init__#s   zExplainWithBackend.__init__rcCs2t||j|jg|j\}|_|_}|_|||Sr )rrjrnrmr)rwrr:_rrr__call__+s zExplainWithBackend.__call__returncCs(t|j}t|j||d|j|j}|S)Nr )rrjrirmrn)rwrkrxrrrrx1s zExplainWithBackend.outputN) rRrSrTrrrr"rrrirxrrrrr s r)- dataclassesrK importlibrtypingrrrrfunctorch.compilerrtorch._functorchrrItorch._functorch.compilersr commonr rr register_backendrrr,r0r5r;r?rArMrNrrPrVr"rr]rarcrh dataclassrirrrrrrsr                   0