o h @sUdZddlZddlZddlZddlZddlZddlZddlZddlm Z ddl m Z m Z m Z mZddlmZddlZddlmZmZmZmZddlmZddlmZmZmZdd lmZmZdd l m!Z!dd l"m#Z#ej$j%Z&ej$j'Z(ed d dZ)edZ*dLddZ+dMddZ,Gdddej-j!Z.ddZ/ej01ddZ2ej01ddZ3ej01ddZ4ej5dNddZ6ej7dfdd Z8d!d"Z9d#d$Z:Gd%d&d&e;ZGd)d*d*e?Z@e@Aej$Bd+d,ZCd-d.ZDd/d0ZEd1d2ZFddd dd3d ddedd f d4d5ZGGd6d7d7eHe ZIGd8d9d9eHe ZJddd dd3d ddedd f d:d;ZKdZLe eMe e feNd<<dd dd3d ddedd f d=d>ZOd?d@ZPGdAdBdBeZQGdCdDdDeQZRdEe e*e)fdFe e*e)ffdGdHZS I    dOdJdKZTdS)PzTracing. This module contains functionality to support the JIT's tracing frontend, notably: * torch.jit.trace * torch.jit.trace_module This is not intended to be imported directly; please use the exposed functionalities in `torch.jit`. N)Enum)AnyCallableOptionalTypeVar) ParamSpec) _get_model_id_qualified_nameget_callable_argument_names is_scripting)function)_CachedForwardscript ScriptModule)_enabled _python_cu)Module)default_tolerancesRT) covariantPcsfdd}|S)Ncst}|s tdd}|dkr%|j}|std|d7}|dks|j}|D]\}}t|tjrF||urF|dkrB|SdSq,dS)Nzfailed to inspect framerrzfailed to get frameself) inspect currentframe RuntimeErrorf_backf_localsitems isinstancetorchTensor)varframeirkv frames_upd/var/www/html/construction_image-detection-poc/venv/lib/python3.10/site-packages/torch/jit/_trace.py_get_interpreter_name_for_var-s   zI_create_interpreter_name_lookup_fn.._get_interpreter_name_for_varr*)r)r,r*r(r+"_create_interpreter_name_lookup_fn,s r-FcCsh|jdd}t|}t}|D]\}}t||vrq|t||r+|||<q|||<q|S)NT keep_vars) state_dicttypesetridadddetach)moduler/r0 filtered_dictseen_idsr&r'r*r*r+_unique_state_dictCs    r9cs6eZdZ    dfdd ZdejfddZZS) ONNXTracedModuleTFcs,t||_||_||_||_||_dSN)super__init__innerstrict_force_outplace_return_inputs_return_inputs_states)rr>r?force_outplace return_inputsreturn_inputs_states __class__r*r+r=Vs  zONNXTracedModule.__init__argscst|\ttdd}gggfdd}tj||tjj \}}j r>|ddfSj rJ|ddfS|dfS)NTr.csg}ttD]}t||tjstd|||qt|}jr2t dd|Dj r=t|j |j rPd|fd<t \}}t|dkr`|dSt |S)NzExpected Tensor argumentcss|] }|jtjdVqdS) memory_formatN)cloner!preserve_format).0xr*r*r+ }sz.wrapper..rr) rangelenr r!r"rappend _unflattenrAtuplerBr>_flatten)rHin_argsr% trace_inputsout_vars_in_descin_vars inputs_statesouts ret_inputsrr*r+wrapperrs&   z)ONNXTracedModule.forward..wrapperr) rUlistr9valuesr!_C_create_graph_by_tracingr-r?r@rArB)rrH module_stater`graph_outr*rZr+forwardhs$  zONNXTracedModule.forward)TFFF)__name__ __module__ __qualname__r=r!r"rh __classcell__r*r*rFr+r:Usr:cs$fddtjdddd|S)Ncsd|durdSt|tjr+|j|jrdntjd|j}|j dur)|j |_ |S|jtjdS)NrI) r r!r"r5rK is_mkldnnrLrequires_grad_ requires_gradgrad)ar' clone_inputr*r+rss   z"_clone_inputs..clone_inputcSs t|tjSr;r r!r"rNr*r*r+s z_clone_inputs..tensors) condition_msg)r _nested_map)rHr*rrr+ _clone_inputss  rzPYTORCH_JIT_TIMEPYTORCH_JIT_DISABLEPYTORCH_JIT_STATSccsts|r tjsdVdStj}tjjdd}tjjdd}||zdVW|||t|d|d| |ddS|||t|d|d| |dw)NT) enable_timing z time: z ms) _JIT_TIMEr!cuda is_availablecurrent_streamEvent record_event synchronizeprint elapsed_time) trace_namenametimestreamstartendr*r*r+_times    $ "rc sttjjs tdttt|ts|f}r!t }d fdd }tj j |dd||dd\}}j |sBJWd n1sLwYrX |||dd \}} t||t|| d S) aB Verify that a JIT compiled model has the same behavior as its uncompiled version along with its backwards pass. If your model returns multiple outputs, you must also specify a `loss_fn` to produce a loss for which the backwards will be computed. This function has side-effects (e.g., it executes your model / saves and loads parameters), so don't expect the model to come out exactly the same as what you passed in. Args: model (compiled torch.nn.Module or function): the module/function to be verified. The module/function definition MUST have been decorated with `@torch.jit.compile`. args (tuple or Tensor): the positional arguments to pass to the compiled function/module to be verified. A non-tuple is assumed to be a single positional argument to be passed to the model. loss_fn (function, optional): the loss function to be applied to the output of the model, before backwards is invoked. By default, we assume that a model returns a single result, and we :func:`torch.sum` before calling backwards; if this is inappropriate, you can pass your own loss function. Note that if a model returns a tuple of results, these are passed as separate positional arguments to `loss_fn`. devices (iterable of device IDs, optional): the GPU devices which the compiled module will be run on. This determines the RNG state we must save when running both compiled and uncompiled versions of the model. zICannot verify an uncompiled module. Add @torch.jit.compile to compile itFcsrtng}t||f\}}}|r||r|j}|}|r.|j|kr.tdt|ts6|f}tj krKt |dkrKt dt |dt|\} }dd| D} |} tj | g|} dd| D} | | fS)Nz#failed to use the compiled functionrzModel returns zO outputs, but default loss function (torch.sum) can only handle a single outputcSg|] }|jtjdqSrIr5rKr!rLrMr'r*r*r+ z/verify..run_fwd_bwd..cSrrrrr*r*r+rr)ra parametersrU clear_cachehitsrr rTr!sumrQ ValueErrorautogradrp)rH force_traceassert_compiledparamsr\rY compiled_fnroutrX saved_outslossgrads saved_grads is_moduleloss_fnmodelr*r+ run_fwd_bwds4  zverify..run_fwd_bwdztorch.jit.verify)_callerT)rN)r)FF)r r!rcCompiledFunction TypeErrorrrTcopydeepcopyr0randomfork_rng has_trace_forload_state_dict _verify_equal) rrHrdevices saved_stateruncompiled_outsuncompiled_grads compiled_outscompiled_gradsr*rr+verifys&#    rcCs6t||D]\}}||dkrtdqdS)Ngư>z!JIT and real computation mismatch)zipsubabsmaxr)xsysrNyr*r*r+r!s rcCsddd|DS)N cSsg|]}d|qS) r*)rMliner*r*r+r(zindent..)join splitlines)sr*r*r+indent'srcseZdZdfdd ZZS)TracingCheckErrorNcsd|_|dur|j|d7_|dur&|jd7_|jt|d7_|dur<|jd7_|jt|d7_t|jdS)NzTracing failed sanity checks! rz+ERROR: Graphs differed across invocations! zERROR: Tensor-valued Constant nodes differed in value across invocations. This often indicates that the tracer has encountered untraceable code. )messagerr<r=)rgraph_diff_errortensor_compare_error extra_msgrFr*r+r=,szTracingCheckError.__init__r;)rirjrkr=rlr*r*rFr+r+src  s|D]} t| tjr | f} |rSi} | D] \} } t| | | <qtjjt|d|| d|||tj dd } | j j | j } t| tjsOt| t rRsR| f} n#retjj|d|||t| dd} ntjj|t| d|||dd} | fdddd fd d }dgfd d fdd}|| d}||| d}|||dr|| d}|||d}tdd|Drt|qdS)N__self__F) check_tracer?r@ _module_class_compilation_unitexample_inputs_is_kwarg _store_inputs)rr?r@rexample_kwarg_inputsr)rr?r@rrc stjj}tj|tj|t|}tdd|}tjj}tj|tj|t|}tdd|}d}||krddl }| | d| d}dt d |d}t||D]Q\}}t|t|kr|d7}| t| dt| d} dt d | d} |} | r| d t | d7} |} | r| d t | d7} || 7}nqfd} t||D]\}}||kr|| fS|d krW|sW|sW|d sq|d d ks|d d krq|d }|d }z tjj||ddWqttfyV}z8| dur!d} | dt t|d7} |}|r>| dt |d7} | dt t|7} WYd}~|| fSd}~wwq|| fS)Nz___torch_mangle_[0-9]+\.rrTz Graph diff: rzFirst diverging operator: z Node diff: zTrace source location: zCheck source location: zprim::Constantvaluet) equal_nanzNode: zSource Location: zComparison exception: )r!rc_jit_pass_canonicalizerf_jit_pass_inline!_jit_pass_erase_shape_informationstrrerdifflibndiffrrrrnodes sourceRangekind mustBeNone hasAttributekindOfrtesting assert_closerAssertionError)mod_canonicalizedmod_strcheck_canonicalized check_strgraph_diff_errorsr graph_diffn_modn_check node_diffsource_printout mod_stack check_stacktensor_compare_errorsmod_tensor_valcheck_tensor_vale compare_stack)check_mod_func traced_funcr*r+graph_diagnostic_info}s               z+_check_trace..graph_diagnostic_infocSst|tr|S|fSr;)r rTrur*r*r+ wrap_retvalsz!_check_trace..wrap_retvalc sz#t|trr|di|}n|t|}dd|D}|WStyH}z\}}d|dtt|}t|||d|d}~ww)NcSsg|] }t|tjr|qSr*rt)rMrr*r*r+rszK_check_trace..run_mod_and_filter_tensor_outputs..z+encountered an exception while running the z with test inputs. Exception: )rr*)r dictrz Exceptionrrr)modinputs running_whatr^rrrmsg)rrrr*r+!run_mod_and_filter_tensor_outputss$ z7_check_trace..run_mod_and_filter_tensor_outputscsdrdSdd<ddjD}t|dkr>d}|d7}|ddd|Ddd 7}|d 7}tj|td d dSdS) NrTcSsg|]}|r|qSr*)isNondeterministicrMopr*r*r+rs zE_check_trace..maybe_warn_nondeterministic..z"Trace had nondeterministic nodes. z2Did you forget call .eval() on your model? Nodes: rcSsg|]}tt|qSr*)rrrr*r*r+rszp This may cause errors in trace checking. To disable trace checking, pass check_trace=False to torch.jit.trace()category stacklevel)rfrrQrwarningswarn TracerWarning) nondeterm_opsnondeterministic_ops_warning) has_warnedrr*r+maybe_warn_nondeterministics*  z1_check_trace..maybe_warn_nondeterministicc sd}tt||D]\}\}}z|jr|}|jr|}|jr%|}|jr,|}|s4|rMtjj | tj | tj t ||dddnl|j sS|j rhtjj ||t ||dddnQt|ddstt|ddrt|ddt|ddksJt||D]\}}tjj ||t ||dddqntjj ||t ||dddWq ty} z#tjdt|dd|dt| tdd d }WYd} ~ q d} ~ ww|S) NTr)rtolatolr is_nestedz Output nr zH. of the traced function does not match the corresponding output of the z. Detailed error: r F) enumerater is_quantized dequantizermto_dense is_complexr!rrtocdoubleris_mpsfloatgetattrunbinddoublerr rrr) original reference match_whatall_okr%origreft_origt_refr)check_tolerancerr*r+compare_outputss            z%_check_trace..compare_outputstracezPython functionzrepeated tracecss|]}|duVqdSr;r*)rMinfor*r*r+rOMsz_check_trace..)r r!r"rrzjit trace_moduler!rcCompilationUnit_c _get_methodrrr.anyr) check_inputsfuncrr,r?rCis_trace_modulerrr copied_dictrdata check_modrr- traced_outsfn_outs check_outs diag_infor*)rr,rrrrrrr+ _check_trace>s      V D   r@c@seZdZeddZdS)rcCs tjdtddtdddS)Nignorez torch.(?!jit))r r6ztorch::jit::fuser::cuda)r filterwarningsrr*r*r*r+ignore_lib_warningsRsz!TracerWarning.ignore_lib_warningsN)rirjrk staticmethodrCr*r*r*r+rQsrcCs,t|tjtfr |fSt|tst|S|Sr;)r r!r"rrT)example_inputsr*r*r+ make_tupleas  rFcCsRt|tr|Stj|rtjjj}tjjj||dddS|dur#t }|||dS)NFT share_types is_tracingr) r rr! _jit_internalmodule_has_exportsr0 _recursive make_stubs_from_exported_methodscreate_script_moduleTopLevelTracedModule)rrrinfer_methods_stubs_fnr*r*r+ make_modulejs    rRcCs|durdSdd|DS)NcSsg|]}d|iqS)rhr*)rMcr*r*r+r|rz%wrap_check_inputs..r*)r6r*r*r+wrap_check_inputsxsrTcCsddlmm}||}||}t||D]K\}}|j|jkr%dS|jtjkr.dSt|t|kr9dSt |tj j rCdSt |tj r\|j |j krRdSt||s[dSq||krcdSqdS)NrFT)torch.utils._pytreeutils_pytree tree_leavesrlayoutr!_mkldnnr1r _subclasses FakeTensorr"dtypeallclose)exportr.pytree flat_export flat_tracer(loadedr*r*r+$analyze_ts_result_with_export_results,       rdgh㈵>c  Cst|tjjrtd|St|tjjr;|dur%t| tr!| }nt dt |d|id|t |||||t| t| d St |drtt|j tjjrt|jdkrt|dur]t| trY| }nt dt |j d|id|t |||||t| t| d St|tjtfr| dur|f}n | durt|tst|}td} t |drt|j tjjrtdt|} t| tr| }tj| || | ||t|}ntj| ||| ||t|}|r|durt||||||d|t| td nt|g|||||d|t| td ||_|S) Nz`The input to trace is already a ScriptModule, tracing it is a no-op. Returning the object as is.z%example_kwarg_inputs should be a dictrh)rrrrzVtrace doesn't support compiling individual module's functions. Please use trace_moduleFr)r r!r0rr rnnrrrr1rThasattrrrir"rTr-AttributeErrorr rc%_create_function_from_trace_with_dictr _create_function_from_tracer@_torchdynamo_inline)r7rEoptimizerr6r,r?r@rrrr var_lookup_fnrtracedr*r*r+ _trace_impls        roc@s&eZdZdZdZdZdefddZdS) _ExportType DIRECT_EXPORTTRACE_AND_EXPORTSOURCE_TO_SOURCEreturncC|jSr;rrr*r*r+__str__+z_ExportType.__str__N)rirjrkrqrrrsrrxr*r*r*r+rp&s rpc@s*eZdZdZdZdZdZdefddZdS) _ExportOutcomeSUCCESSFAILED_TO_EXPORT FAILED_TO_RUNACCURACY_ERRORrtcCrur;rvrwr*r*r+rx5ryz_ExportOutcome.__str__N) rirjrkr{r|r}r~rrxr*r*r*r+rz/s rzc sts|S|durtjdtddddlm} mm} t|||||||||| | | }| dt |d| rdd l m dd l m }m}t|||d |||||| | | d }||| \}}fd d}dd}fdd}t|tjjsx||||tj||||tj||||tj|S)a Trace a function and return an executable or :class:`ScriptFunction` that will be optimized using just-in-time compilation. Tracing is ideal for code that operates only on ``Tensor``\\s and lists, dictionaries, and tuples of ``Tensor``\\s. Using `torch.jit.trace` and `torch.jit.trace_module`, you can turn an existing module or Python function into a TorchScript :class:`ScriptFunction` or :class:`ScriptModule`. You must provide example inputs, and we run the function, recording the operations performed on all the tensors. * The resulting recording of a standalone function produces `ScriptFunction`. * The resulting recording of `nn.Module.forward` or `nn.Module` produces `ScriptModule`. This module also contains any parameters that the original module had as well. Warning: Tracing only correctly records functions and modules which are not data dependent (e.g., do not have conditionals on data in tensors) and do not have any untracked external dependencies (e.g., perform input/output or access global variables). Tracing only records operations done when the given function is run on the given tensors. Therefore, the returned `ScriptModule` will always run the same traced graph on any input. This has some important implications when your module is expected to run different sets of operations, depending on the input and/or the module state. For example, * Tracing will not record any control-flow like if-statements or loops. When this control-flow is constant across your module, this is fine and it often inlines the control-flow decisions. But sometimes the control-flow is actually part of the model itself. For instance, a recurrent network is a loop over the (possibly dynamic) length of an input sequence. * In the returned :class:`ScriptModule`, operations that have different behaviors in ``training`` and ``eval`` modes will always behave as if it is in the mode it was in during tracing, no matter which mode the `ScriptModule` is in. In cases like these, tracing would not be appropriate and :func:`scripting ` is a better choice. If you trace such models, you may silently get incorrect results on subsequent invocations of the model. The tracer will try to emit warnings when doing something that may cause an incorrect trace to be produced. Args: func (callable or torch.nn.Module): A Python function or `torch.nn.Module` that will be run with `example_inputs`. `func` arguments and return values must be tensors or (possibly nested) tuples that contain tensors. When a module is passed `torch.jit.trace`, only the ``forward`` method is run and traced (see :func:`torch.jit.trace ` for details). Keyword arguments: example_inputs (tuple or torch.Tensor or None, optional): A tuple of example inputs that will be passed to the function while tracing. Default: ``None``. Either this argument or ``example_kwarg_inputs`` should be specified. The resulting trace can be run with inputs of different types and shapes assuming the traced operations support those types and shapes. `example_inputs` may also be a single Tensor in which case it is automatically wrapped in a tuple. When the value is None, ``example_kwarg_inputs`` should be specified. check_trace (``bool``, optional): Check if the same inputs run through traced code produce the same outputs. Default: ``True``. You might want to disable this if, for example, your network contains non- deterministic ops or if you are sure that the network is correct despite a checker failure. check_inputs (list of tuples, optional): A list of tuples of input arguments that should be used to check the trace against what is expected. Each tuple is equivalent to a set of input arguments that would be specified in ``example_inputs``. For best results, pass in a set of checking inputs representative of the space of shapes and types of inputs you expect the network to see. If not specified, the original ``example_inputs`` are used for checking check_tolerance (float, optional): Floating-point comparison tolerance to use in the checker procedure. This can be used to relax the checker strictness in the event that results diverge numerically for a known reason, such as operator fusion. strict (``bool``, optional): run the tracer in a strict mode or not (default: ``True``). Only turn this off when you want the tracer to record your mutable container types (currently ``list``/``dict``) and you are sure that the container you are using in your problem is a ``constant`` structure and does not get used as control flow (if, for) conditions. example_kwarg_inputs (dict, optional): This parameter is a pack of keyword arguments of example inputs that will be passed to the function while tracing. Default: ``None``. Either this argument or ``example_inputs`` should be specified. The dict will be unpacking by the arguments name of the traced function. If the keys of the dict don't not match with the traced function's arguments name, a runtime exception will be raised. Returns: If `func` is `nn.Module` or ``forward`` of `nn.Module`, `trace` returns a :class:`ScriptModule` object with a single ``forward`` method containing the traced code. The returned `ScriptModule` will have the same set of sub-modules and parameters as the original ``nn.Module``. If ``func`` is a standalone function, ``trace`` returns `ScriptFunction`. Example (tracing a function): .. testcode:: import torch def foo(x, y): return 2 * x + y # Run `foo` with the provided inputs and record the tensor operations traced_foo = torch.jit.trace(foo, (torch.rand(3), torch.rand(3))) # `traced_foo` can now be run with the TorchScript interpreter or saved # and loaded in a Python-free environment Example (tracing an existing module):: import torch import torch.nn as nn class Net(nn.Module): def __init__(self) -> None: super().__init__() self.conv = nn.Conv2d(1, 1, 3) def forward(self, x): return self.conv(x) n = Net() example_weight = torch.rand(1, 1, 3, 3) example_forward_input = torch.rand(1, 1, 3, 3) # Trace a specific method and construct `ScriptModule` with # a single `forward` method module = torch.jit.trace(n.forward, example_forward_input) # Trace a module (implicitly traces `forward`) and construct a # `ScriptModule` with a single `forward` method module = torch.jit.trace(n, example_forward_input) N^`optimize` is deprecated and has no effect. Use `with torch.jit.optimized_execution()` insteadr r)check_if_torch_exportable!log_torch_jit_trace_exportabilitylog_torchscript_usager.)model_idTS2EPConverter)"_convert_ts_to_export_experimental$_process_jit_trace_inputs_for_exportF) rErlrr6r,r?r@rrrrc s4z||}Wn!ty'}z|}dt|ttjdWYd}~dSd}~wwz|||}Wn!tyP}zdt|ttjt|WYd}~dSd}~wwz||}Wn!tyx}zdt|ttjt|WYd}~dSd}~wwt||sdt|ttjddSdt|ttjddS)Nr. succeededzaccuracy error)rrrzr{r|r}rdr~) func_to_export export_func export_args export_type traced_resultrrY ep_moduler_)rr*r+_log_exportabilitysT   z!trace.._log_exportabilitycSstjj||ddS)NF)r?)r!r_r6r7rr*r*r+_direct_export_and_lower>sz'trace.._direct_export_and_lowercs||Sr;)convertr6rrr*r+&_convert_ts_to_export_source_to_sourceAsz5trace.._convert_ts_to_export_source_to_source)rr r FutureWarningtorch._utils_internalrrrrortorch._export.converterrtorch.export._tracerrr r!r0rrprqrrrs)r7rErlrr6r,r?r@rrrrrrrrrtraced_func_for_exportrrYrrrr*)rrr+r.9s"  * r._trace_module_mapc  sts|S|durtjdtddtd} t|tjjst dt|t s(t dtj j j } zifdd |d <tj j _ |d t||| }|D]\}}|d krc|}t||}t|}n t||}t|}t|t r| r|D]}||vrd d |d}td|d|qu|j|||| |||| nt|}|j|||| |||| |j|}|r|durt||||||d|| d qOt|g|||||d|| d qOW| tj j _ |S| tj j _ w)a Trace a module and return an executable :class:`ScriptModule` that will be optimized using just-in-time compilation. When a module is passed to :func:`torch.jit.trace `, only the ``forward`` method is run and traced. With ``trace_module``, you can specify a dictionary of method names to example inputs to trace (see the ``inputs``) argument below. See :func:`torch.jit.trace ` for more information on tracing. Args: mod (torch.nn.Module): A ``torch.nn.Module`` containing methods whose names are specified in ``inputs``. The given methods will be compiled as a part of a single `ScriptModule`. inputs (dict): A dict containing sample inputs indexed by method names in ``mod``. The inputs will be passed to methods whose names correspond to inputs' keys while tracing. ``{ 'forward' : example_forward_input, 'method2': example_method2_input}`` Keyword arguments: check_trace (``bool``, optional): Check if the same inputs run through traced code produce the same outputs. Default: ``True``. You might want to disable this if, for example, your network contains non- deterministic ops or if you are sure that the network is correct despite a checker failure. check_inputs (list of dicts, optional): A list of dicts of input arguments that should be used to check the trace against what is expected. Each tuple is equivalent to a set of input arguments that would be specified in ``inputs``. For best results, pass in a set of checking inputs representative of the space of shapes and types of inputs you expect the network to see. If not specified, the original ``inputs`` are used for checking check_tolerance (float, optional): Floating-point comparison tolerance to use in the checker procedure. This can be used to relax the checker strictness in the event that results diverge numerically for a known reason, such as operator fusion. example_inputs_is_kwarg (``bool``, optional): This parameter indicate whether the example inputs is a pack pack of keyword arguments. Default: ``False``. Returns: A :class:`ScriptModule` object with a single ``forward`` method containing the traced code. When ``func`` is a ``torch.nn.Module``, the returned :class:`ScriptModule` will have the same set of sub-modules and parameters as ``func``. Example (tracing a module with multiple methods):: import torch import torch.nn as nn class Net(nn.Module): def __init__(self) -> None: super().__init__() self.conv = nn.Conv2d(1, 1, 3) def forward(self, x): return self.conv(x) def weighted_kernel_sum(self, weight): return weight * self.conv.weight n = Net() example_weight = torch.rand(1, 1, 3, 3) example_forward_input = torch.rand(1, 1, 3, 3) # Trace a specific method and construct `ScriptModule` with # a single `forward` method module = torch.jit.trace(n.forward, example_forward_input) # Trace a module (implicitly traces `forward`) and construct a # `ScriptModule` with a single `forward` method module = torch.jit.trace(n, example_forward_input) # Trace specific methods on a module (specified in `inputs`), constructs # a `ScriptModule` with `forward` and `weighted_kernel_sum` methods inputs = {"forward": example_forward_input, "weighted_kernel_sum": example_weight} module = torch.jit.trace_module(n, inputs) Nrrrrz.expected torch.nn.Module as the first argumentz3expected a dictionary of (method_name, input) pairscs4|D]\}}|d|}||<||qdS)N.)named_children)rprefixrchildsubmod_qualnameregister_submodstrace_module_mapr*r+rs   z&trace_module..register_submods__modulerh[,]'z\' is not in forward() method's arguments, valid arguments name are Tre)rr rrr-r r!rfrrhrr0_tracerrRrr!r r NameErrorr3#_create_method_from_trace_with_dictrF_create_method_from_tracer4r@)rrrlrr6r,r?r@rrrrrmold_module_mapr6 method_namerEr7forward_methodargument_nameskeyvalid_argumentscheck_trace_methodr*rr+r1`s\            I r1cCstrdStjS)zReturn a boolean value. Returns ``True`` in tracing (if a function is called during the tracing of code with ``torch.jit.trace``) and ``False`` otherwise. F)r r!rc _is_tracingr*r*r*r+rI,s rIcsReZdZdZdfdd ZddZfddZfd d Zd d Zd dZ Z S) TracedModuleTNc stt|tjjsJtGdddtjj}tjt ||_ |}fdd}|j |_ |j D]\}}|durG||j |<||q6|j D]\}} | dur^| |j|<|| qM|j D]\}} tj| r~||j vr~||jvr~t||| qd|jrtdt||j D]\}} | durqt| tdd|j|<qtjjj|ddd d d } t |j|jd <| |jd <dD]}t||qdS)Nc@s eZdZdS)z.TracedModule.__init__..QualnameWrapperN)rirjrkr*r*r*r+QualnameWrapperGsrcs|vrtd|dS)Nz=TracedModules don't support parameter sharing between modules)rr4)paramid_setr*r+ check_uniquePs z+TracedModule.__init__..check_uniquez=Modules that have backward hooks assigned can't be compiled: rJcSsdS)Nr*r*)r6r*r*r+rvwsz'TracedModule.__init__..FTrG_name_actual_script_module) _parameters_buffers_modulestraining)r<r=r r!rfrr2rKr r1_jit_override_qualnamerrrr__dict__rc_jit_is_script_objectsetattr_backward_hooksrrrrRrr0rMrOridelattr) rr(rrr tmp_modulerrrbufval submodule script_modulerFrr+r=:s^           zTracedModule.__init__cOstd)Nz"Trace submodules cannot be called.)r)rrHkwargsr*r*r+rhszTracedModule.forwardcs"d|jvr t|St|j|SNr)rr< __getattr__r!r)rattrrFr*r+rs   zTracedModule.__getattr__cs*d|jvr t||St|j||dSr)rr< __setattr__rr)rrrrFr*r+rs zTracedModule.__setattr__cCrur;rrwr*r*r+ _get_nameryzTracedModule._get_namecCs d|jS)Nzoriginal_name=rrwr*r*r+ extra_reprs zTracedModule.extra_repr)NN) rirjrk_disable_script_metar=rhrrrrrlr*r*rFr+r7sE  rc@s,eZdZUeZedefed<ddZdS)rP.rhcCs|jd|dS)z Re-construct an instance of TopLevelTracedModule using an instance of a C++ module. Args: cpp_module: The C++ module that this TopLevelTracedModule will be rebuilt around. rN)r _reconstruct)r cpp_moduler*r*r+rsz!TopLevelTracedModule._reconstructN) rirjrkr rhrr__annotations__rr*r*r*r+rPs  rPfnrtcs:tdtjdtjdtffdd _d_S)NrHrrtcs,ts |i|Stj}||i|Sr;)rIr __original_fn)rHrrrr`r*r+r`s z#_script_if_tracing..wrapperT) functoolswrapsrrHrrr__script_if_tracing_wrapper)rr*rr+_script_if_tracings "rr*cCs:|duri}t|ts|f}t||||||i|}|S)aReturn a tuple on tracing a function or model. .. warning:: This function is internal-only and should only be used by the ONNX exporter. If you are trying to get a graph through tracing, please go through the public API instead:: trace = torch.jit.trace(nn.LSTMCell(), (input, hidden)) trace_graph = trace.graph Trace a function or model, returning a tuple consisting of the both the *trace* of an execution, as well as the original return value. If return_inputs, also returns the trace inputs as part of the tuple Tracing is guaranteed not to change the semantics of the function/module that is traced. Args: f (torch.nn.Module or function): the function or module to be traced. args (tuple or Tensor): the positional arguments to pass to the function/module to be traced. A non-tuple is assumed to be a single positional argument to be passed to the model. kwargs (dict): the keyword arguments to pass to the function/module to be traced. Example (trace a cell): .. testcode:: trace = torch.jit.trace(nn.LSTMCell(), (input, hidden)) N)r rTr:)frHrr?r@rDrBr^r*r*r+_get_trace_graphs)  r)r)F)T)r*NTFFF)U__doc__ contextlibrrrosrr enumrtypingrrrrtyping_extensionsrr!torch._jit_internalrr r r torch.autogradr torch.jit._scriptr rrtorch.jit._staterrtorch.nnrtorch.testing._comparisonrrc _jit_flattenrU_jit_unflattenrSrrr-r9rfr:rzenvirongetr _JIT_DISABLE _JIT_STATScontextmanagerrrrrrrrno_gradr@WarningrrC_tracer_warn_use_pythonrFrRrTrdrorrprzr.rrrr1rIrrPrrr*r*r*r+s         E Z        & M \"