U KcHt @sXddlmZmZmZmZmZmZddlZddlZddl m Z ddl m Z ddl mZddl mZddlmZdd lmZmZdd lmZdd lmZmZmZdd lmZdd lmZddlmZddlm Z ej!j"ddddZ#ej!j"ddddZ$d9e e%eeee&efdfeeee&efdfe dddZ'Gddde(Z)Gddde(Z*d:ej!j"ee ee&effe%eedfeeee&efdfeee ee&effeeee&efdfe%ed dd Z+d;ej!j"ee ee&effe%eedfeeee&efdfeeee&efdfe d!d"d#Z,dej!j"ee ee&effeedfeeee&efdfeeee&efdfed*d+d,Z/d?e e%eeee&efdfe%e%ee ee&efdfeeee&efdfej!j"d.d/d0Z0d@e eeee&efdfe%ee ee&efdfeeee&efdfej!j"d1d2d3Z1dAe eeee&efdfe%ee ee&efdfeeee&efdfej!j"d1d4d5Z2dBe e%eeee&efdfej!j"d6d7d8Z3dS)C)AnyDictOptionalSetTupleUnionN) GraphModule)QuantizationTracer)fuse)prepare)convert) BackendConfigget_tensorrt_backend_config)ObservedGraphModule)ConvertCustomConfigFuseCustomConfigPrepareCustomConfig)graph_pretty_str)get_custom_module_class_keys)#get_skipped_module_name_and_classes)QConfigMapping)modelreturncCs*t|ts&tdtt|dddS)Nz,input model must be a GraphModule, Got type:z Please make zsure to follow the tutorials.) isinstancer ValueErrorstrtype)rrE/tmp/pip-unpacked-wheel-gikjz4vx/torch/ao/quantization/quantize_fx.py_check_is_graph_modules  r cCsfg}|D].\}}t|tjjjjr2||q t|q |D] }|j |=tjjj |j |<q@dS)z1 Swap FloatFunctional with FXFloatFunctional N) Znamed_childrenrtorchZaonn quantizedZFloatFunctionalappend_swap_ff_with_fxffZ_modulesZFXFloatFunctional)rZmodules_to_swapnamemodulerrrr%$s  r%) graph_moduleis_qatfuse_custom_configbackend_configrcCst|t||||S)z Internal helper function to fuse modules in preparation for quantization Args: graph_module: GraphModule object from symbolic tracing (torch.fx.symbolic_trace) )r r )r(r)r*r+rrr_fuse_fx3s r,cs(eZdZdZeedfdd ZZS)Scopea/ Scope object that records the module path and the module type of a module. Scope is used to track the information of the module that contains a Node in a Graph of GraphModule. For example:: class Sub(torch.nn.Module): def forward(self, x): # This will be a call_method Node in GraphModule, # scope for this would be (module_path="sub", module_type=Sub) return x.transpose(1, 2) class M(torch.nn.Module): def __init__(self): self.sub = Sub() def forward(self, x): # This will be a call_method Node as well, # scope for this would be (module_path="", None) x = x.transpose(1, 2) x = self.sub(x) return x ) module_path module_typecst||_||_dSN)super__init__r.r/)selfr.r/ __class__rrr2[s zScope.__init__)__name__ __module__ __qualname____doc__rrr2 __classcell__rrr4rr-Csr-cs>eZdZdZeejjedfdd Z ddZ ddZ Z S) ScopeContextManagerz A context manager to track the Scope of Node during symbolic tracing. When entering a forward function of a Module, we'll update the scope information of the current module, and when we exit, we'll restore the previous scope information. )scopecurrent_modulecurrent_module_pathcs8t|j|_|j|_||_||j_t||j_dSr0)r1r2r/prev_module_typer.prev_module_pathr<r)r3r<r=r>r4rrr2gs  zScopeContextManager.__init__cCsdSr0r)r3rrr __enter__qszScopeContextManager.__enter__cGs|j|j_|j|j_dSr0)r@r<r.r?r/)r3argsrrr__exit__ts  zScopeContextManager.__exit__) r6r7r8r9r-r!r"Modulerr2rArCr:rrr4rr;as r;F.) rqconfig_mappingr)example_inputsprepare_custom_config_equalization_configr+is_standalone_modulerc Cs|dkrt}|dkrt}t|tr:tdt|}t|t||\}} |j } t || } t || |} | D]} t | | t|| qtt|j }t| |||} t| ||| j|||||d }| D]} t || t|| q|S)a7 Internal helper function for prepare_fx Args: `model`, `qconfig_mapping`, `prepare_custom_config`, `_equalization_config`: see docs for :func:`~torch.ao.quantization.prepare_fx` `is_standalone_module`: a boolean flag indicates whether we are quantizing a standalone module or not, a standalone module is a submodule of the parent module that is not inlined in the forward graph of the parent module, the way we quantize standalone module is described in: :func:`~torch.ao.quantization._prepare_standalone_module_fx` NzPassing a prepare_custom_config_dict to prepare is deprecated and will not be supported in a future version. Please pass in a PrepareCustomConfig instead.)rFrGrHr+rI)rrrrwarningswarn from_dictr%rpreserved_attributesr rtracesetattrgetattrrZset_preserved_attributesr,r Znode_name_to_scope)rrEr)rFrGrHr+rIZskipped_module_namesZskipped_module_classesrMZtracerr( attr_namer*preparedrrr _prepare_fxzsL    rS)rrEr)rFrGr+rc Cst||||||ddS)a [Internal use only] Prepare a standalone module, so that it can be used when quantizing the parent module. standalone_module means it a submodule that is not inlined in parent module, and will be quantized separately as one unit. How the standalone module is observed is specified by `input_quantized_idxs` and `output_quantized_idxs` in the prepare_custom_config for the standalone module Returns: * model(GraphModule): prepared standalone module. It has these attributes: * `_standalone_module_input_quantized_idxs(List[Int])`: a list of indexes for the graph input that is expected to be quantized, same as input_quantized_idxs configuration provided for the standalone module * `_standalone_module_output_quantized_idxs(List[Int])`: a list of indexs for the graph output that is quantized same as input_quantized_idxs configuration provided for the standalone module T)r+rI)rS)rrEr)rFrGr+rrr_prepare_standalone_module_fxsrT)rr*r+rcCs|dkrt}t|tr,tdt|}tjdtj |}t }|rXt |j }|D]}t ||t||q\t|d||S)a  Fuse modules like conv+bn, conv+bn+relu etc, model must be in eval mode. Fusion rules are defined in torch.quantization.fx.fusion_pattern.py Args: * `model` (torch.nn.Module): a torch.nn.Module model * `fuse_custom_config` (FuseCustomConfig): custom configurations for fuse_fx. See :class:`~torch.ao.quantization.fx.custom_config.FuseCustomConfig` for more details Example:: from torch.ao.quantization import fuse_fx m = Model().eval() m = fuse_fx(m) NzPassing a fuse_custom_config_dict to fuse is deprecated and will not be supported in a future version. Please pass in a FuseCustomConfig instead.z$quantization_api.quantize_fx.fuse_fxF)rrrrJrKrLr!_C_log_api_usage_oncefxZsymbolic_tracesetrMrOrPr,)rr*r+r(rMrQrrrfuse_fxs     rY)rrErFrGrHr+rcCs tjdt||d||||S)a Prepare a model for post training static quantization Args: * `model` (torch.nn.Module): torch.nn.Module model * `qconfig_mapping` (QConfigMapping): QConfigMapping object to configure how a model is quantized, see :class:`~torch.ao.quantization.qconfig_mapping.QConfigMapping` for more details * `example_inputs` (Tuple[Any, ...]): Example inputs for forward function of the model, Tuple of positional args (keyword args can be passed as positional args as well) * `prepare_custom_config` (PrepareCustomConfig): customization configuration for quantization tool. See :class:`~torch.ao.quantization.fx.custom_config.PrepareCustomConfig` for more details * `_equalization_config`: config for specifying how to perform equalization on the model * `backend_config` (BackendConfig): config that specifies how operators are quantized in a backend, this includes how the operaetors are observed, supported fusion patterns, how quantize/dequantize ops are inserted, supported dtypes etc. See :class:`~torch.ao.quantization.backend_config.BackendConfig` for more details Return: A GraphModule with observer (configured by qconfig_mapping), ready for calibration Example:: import torch from torch.ao.quantization import get_default_qconfig_mapping from torch.ao.quantization import prepare_fx class Submodule(torch.nn.Module): def __init__(self): super().__init__() self.linear = torch.nn.Linear(5, 5) def forward(self, x): x = self.linear(x) return x class M(torch.nn.Module): def __init__(self): super().__init__() self.linear = torch.nn.Linear(5, 5) self.sub = Submodule() def forward(self, x): x = self.linear(x) x = self.sub(x) + x return x # initialize a floating point model float_model = M().eval() # define calibration function def calibrate(model, data_loader): model.eval() with torch.no_grad(): for image, target in data_loader: model(image) # qconfig is the configuration for how we insert observers for a particular # operator # qconfig = get_default_qconfig("fbgemm") # Example of customizing qconfig: # qconfig = torch.ao.quantization.QConfig( # activation=MinMaxObserver.with_args(dtype=torch.qint8), # weight=MinMaxObserver.with_args(dtype=torch.qint8)) # `activation` and `weight` are constructors of observer module # qconfig_mapping is a collection of quantization configurations, user can # set the qconfig for each operator (torch op calls, functional calls, module calls) # in the model through qconfig_mapping # the following call will get the qconfig_mapping that works best for models # that target "fbgemm" backend qconfig_mapping = get_default_qconfig_mapping("fbgemm") # We can customize qconfig_mapping in different ways. # e.g. set the global qconfig, which means we will use the same qconfig for # all operators in the model, this can be overwritten by other settings # qconfig_mapping = QConfigMapping().set_global(qconfig) # e.g. quantize the linear submodule with a specific qconfig # qconfig_mapping = QConfigMapping().set_module_name("linear", qconfig) # e.g. quantize all nn.Linear modules with a specific qconfig # qconfig_mapping = QConfigMapping().set_object_type(torch.nn.Linear, qconfig) # for a more complete list, please see the docstring for :class:`torch.ao.quantization.QConfigMapping` # argument # example_inputs is a tuple of inputs, that is used to infer the type of the # outputs in the model # currently it's not used, but please make sure model(*example_inputs) runs example_inputs = (torch.randn(1, 3, 224, 224),) # TODO: add backend_config after we split the backend_config for fbgemm and qnnpack # e.g. backend_config = get_default_backend_config("fbgemm") # `prepare_fx` inserts observers in the model based on qconfig_mapping and # backend_config. If the configuration for an operator in qconfig_mapping # is supported in the backend_config (meaning it's supported by the target # hardware), we'll insert observer modules according to the qconfig_mapping # otherwise the configuration in qconfig_mapping will be ignored # # Example: # in qconfig_mapping, user sets linear module to be quantized with quint8 for # activation and qint8 for weight: # qconfig = torch.ao.quantization.QConfig( # observer=MinMaxObserver.with_args(dtype=torch.quint8), # weight=MinMaxObserver.with-args(dtype=torch.qint8)) # Note: current qconfig api does not support setting output observer, but # we may extend this to support these more fine grained control in the # future # # qconfig_mapping = QConfigMapping().set_object_type(torch.nn.Linear, qconfig) # in backend config, linear module also supports in this configuration: # weighted_int8_dtype_config = DTypeConfig( # input_dtype=torch.quint8, # output_dtype=torch.quint8, # weight_dtype=torch.qint8, # bias_type=torch.float) # linear_pattern_config = BackendPatternConfig(torch.nn.Linear) \ # .set_observation_type(ObservationType.OUTPUT_USE_DIFFERENT_OBSERVER_AS_INPUT) \ # .add_dtype_config(weighted_int8_dtype_config) \ # ... # backend_config = BackendConfig().set_backend_pattern_config(linear_pattern_config) # `prepare_fx` will check that the setting requested by suer in qconfig_mapping # is supported by the backend_config and insert observers and fake quant modules # in the model prepared_model = prepare_fx(float_model, qconfig_mapping, example_inputs) # Run calibration calibrate(prepared_model, sample_inference_data) z'quantization_api.quantize_fx.prepare_fxFr!rUrVrS)rrErFrGrHr+rrr prepare_fx s r[)rrErFrGr+rcCs tjdt||d|||dS)a Prepare a model for quantization aware training Args: * `model` (torch.nn.Module): torch.nn.Module model * `qconfig_mapping` (QConfigMapping): see :func:`~torch.ao.quantization.prepare_fx` * `example_inputs` (Tuple[Any, ...]): see :func:`~torch.ao.quantization.prepare_fx` * `prepare_custom_config` (PrepareCustomConfig): see :func:`~torch.ao.quantization.prepare_fx` * `backend_config` (BackendConfig): see :func:`~torch.ao.quantization.prepare_fx` Return: A GraphModule with fake quant modules (configured by qconfig_mapping and backend_config), ready for quantization aware training Example:: import torch from torch.ao.quantization import get_default_qat_qconfig_mapping from torch.ao.quantization import prepare_fx class Submodule(torch.nn.Module): def __init__(self): super().__init__() self.linear = torch.nn.Linear(5, 5) def forward(self, x): x = self.linear(x) return x class M(torch.nn.Module): def __init__(self): super().__init__() self.linear = torch.nn.Linear(5, 5) self.sub = Submodule() def forward(self, x): x = self.linear(x) x = self.sub(x) + x return x # initialize a floating point model float_model = M().train() # (optional, but preferred) load the weights from pretrained model # float_model.load_weights(...) # define the training loop for quantization aware training def train_loop(model, train_data): model.train() for image, target in data_loader: ... # qconfig is the configuration for how we insert observers for a particular # operator # qconfig = get_default_qconfig("fbgemm") # Example of customizing qconfig: # qconfig = torch.ao.quantization.QConfig( # activation=FakeQuantize.with_args(observer=MinMaxObserver.with_args(dtype=torch.qint8)), # weight=FakeQuantize.with_args(observer=MinMaxObserver.with_args(dtype=torch.qint8))) # `activation` and `weight` are constructors of observer module # qconfig_mapping is a collection of quantization configurations, user can # set the qconfig for each operator (torch op calls, functional calls, module calls) # in the model through qconfig_mapping # the following call will get the qconfig_mapping that works best for models # that target "fbgemm" backend qconfig_mapping = get_default_qat_qconfig("fbgemm") # We can customize qconfig_mapping in different ways, please take a look at # the doctring for :func:`~torch.ao.quantization.prepare_fx` for different ways # to configure this # example_inputs is a tuple of inputs, that is used to infer the type of the # outputs in the model # currently it's not used, but please make sure model(*example_inputs) runs example_inputs = (torch.randn(1, 3, 224, 224),) # TODO: add backend_config after we split the backend_config for fbgemm and qnnpack # e.g. backend_config = get_default_backend_config("fbgemm") # `prepare_qat_fx` inserts observers in the model based on qconfig_mapping and # backend_config, if the configuration for an operator in qconfig_mapping # is supported in the backend_config (meaning it's supported by the target # hardware), we'll insert fake_quantize modules according to the qconfig_mapping # otherwise the configuration in qconfig_mapping will be ignored # see :func:`~torch.ao.quantization.prepare_fx` for a detailed explanation of # how qconfig_mapping interacts with backend_config prepared_model = prepare_qat_fx(float_model, qconfig_mapping, example_inputs) # Run training train_loop(prepared_model, train_loop) z+quantization_api.quantize_fx.prepare_qat_fxT)r+rZ)rrErFrGr+rrrprepare_qat_fxs_ r\T)r( is_referenceconvert_custom_configrI_remove_qconfigrEr+rc Csp|dkrt}t|tr,tdt|}t|t|||||||d}|j}|D]} t || t || qT|S)ze `is_standalone_module`: see docs in :func:`~torch.ao.quantization.prepare_standalone_module_fx` NzPassing a convert_custom_config_dict to convert is deprecated and will not be supported in a future version. Please pass in a ConvertCustomConfig instead.)Z_remove_qconfig_flagrEr+) rrrrJrKrLr r rMrOrP) r(r]r^rIr_rEr+r#rMrQrrr _convert_fx s*    r`)r(r^r_rEr+rcCs tjdt|d||||dS)a Convert a calibrated or trained model to a quantized model Args: * `graph_module` (torch.fx.GraphModule): A prepared and calibrated/trained model (GraphModule) * `convert_custom_config` (ConvertCustomConfig): custom configurations for convert function. See :class:`~torch.ao.quantization.fx.custom_config.ConvertCustomConfig` for more details * `_remove_qconfig` (bool): Option to remove the qconfig attributes in the model after convert. * `qconfig_mapping` (QConfigMapping): config for specifying how to convert a model for quantization. The keys must include the ones in the qconfig_mapping passed to `prepare_fx` or `prepare_qat_fx`, with the same values or `None`. Additional keys can be specified with values set to `None`. For each entry whose value is set to None, we skip quantizing that entry in the model:: qconfig_mapping = QConfigMapping .set_global(qconfig_from_prepare) .set_object_type(torch.nn.functional.add, None) # skip quantizing torch.nn.functional.add .set_object_type(torch.nn.functional.linear, qconfig_from_prepare) .set_module_name("foo.bar", None) # skip quantizing module "foo.bar" * `backend_config` (BackendConfig): A configuration for the backend which describes how operators should be quantized in the backend, this includes quantization mode support (static/dynamic/weight_only), dtype support (quint8/qint8 etc.), observer placement for each operators and fused operators. See :class:`~torch.ao.quantization.backend_config.BackendConfig` for more details Return: A quantized model (torch.nn.Module) Example:: # prepared_model: the model after prepare_fx/prepare_qat_fx and calibration/training # convert_fx converts a calibrated/trained model to a quantized model for the # target hardware, this includes converting the model first to a reference # quantized model, and then lower the reference quantized model to a backend # Currently, the supported backends are fbgemm (onednn), qnnpack (xnnpack) and # they share the same set of quantized operators, so we are using the same # lowering procedure # # backend_config defines the corresponding reference quantized module for # the weighted modules in the model, e.g. nn.Linear # TODO: add backend_config after we split the backend_config for fbgemm and qnnpack # e.g. backend_config = get_default_backend_config("fbgemm") quantized_model = convert_fx(prepared_model) z'quantization_api.quantize_fx.convert_fxFr]r^r_rEr+r!rUrVr`r(r^r_rEr+rrr convert_fx3s8 rdcCs tjdt|d||||dS)a~ Convert a calibrated or trained model to a reference quantized model, see https://github.com/pytorch/rfcs/blob/master/RFC-0019-Extending-PyTorch-Quantization-to-Custom-Backends.md for more details, reference quantzied model is a standard representation of a quantized model provided by FX Graph Mode Quantization, it can be further lowered to run on the target hardware, like accelerators Args: * `graph_module` (GraphModule): A prepared and calibrated/trained model (GraphModule) * `convert_custom_config` (ConvertCustomConfig): custom configurations for convert function. See :func:`~torch.ao.quantization.quantize_fx.convert_fx` for more details. * `_remove_qconfig` (bool): Option to remove the qconfig attributes in the model after convert. * `qconfig_mapping` (QConfigMapping): config for specifying how to convert a model for quantization. See :func:`~torch.ao.quantization.quantize_fx.convert_fx` for more details. * `backend_config` (BackendConfig): A configuration for the backend which describes how operators should be quantized in the backend. See :func:`~torch.ao.quantization.quantize_fx.convert_fx` for more details. Return: A reference quantized model (GraphModule) Example:: # prepared_model: the model after prepare_fx/prepare_qat_fx and calibration/training # TODO: add backend_config after we split the backend_config for fbgemm and qnnpack # e.g. backend_config = get_default_backend_config("fbgemm") reference_quantized_model = convert_to_reference_fx(prepared_model) z4quantization_api.quantize_fx.convert_to_reference_fxTrarbrcrrrconvert_to_reference_fxvs' re)r(r]r^rcCst|||ddS)aw [Internal use only] Convert a model produced by :func:`~torch.ao.quantization.prepare_standalone_module_fx` and convert it to a quantized model Returns a quantized standalone module, whether input/output is quantized is specified by prepare_custom_config, with input_quantized_idxs, output_quantized_idxs, please see docs for prepare_fx for details T)rI)r`)r(r]r^rrr_convert_standalone_module_fxs rf)NN)NNNF)NN)NN)NNN)NN)NFTNN)NTNN)NTNN)FN)4typingrrrrrrrJr!Ztorch.fxrZ fx.tracerr rWr r Z fx.convertr r+rrZfx.graph_modulerZfx.custom_configrrrZfx.utilsrrrrErr"rDr r%boolrr,objectr-r;rSrTrYr[r\r`rdrerfrrrrs               G  + +    m ( E 4