fastdeploy::vision::classification::ResNet Class Reference

Torchvision ResNet series model. More...

#include <resnet.h>

Inheritance diagram for fastdeploy::vision::classification::ResNet:

Collaboration diagram for fastdeploy::vision::classification::ResNet:

Public Member Functions
	ResNet (const std::string &model_file, const std::string &params_file="", const RuntimeOption &custom_option=RuntimeOption(), const ModelFormat &model_format=ModelFormat::ONNX)
	Set path of model file and the configuration of runtime. More...
virtual std::string	ModelName () const
	Get model's name.
virtual bool	Predict (cv::Mat *im, ClassifyResult *result, int topk=1)
	Predict for the input "im", the result will be saved in "result". More...
Public Member Functions inherited from fastdeploy::FastDeployModel
virtual bool	Infer (std::vector< FDTensor > &input_tensors, std::vector< FDTensor > *output_tensors)
	Inference the model by the runtime. This interface is included in the Predict() function, so we don't call Infer() directly in most common situation.
virtual bool	Infer ()
	Inference the model by the runtime. This interface is using class member reused_input_tensors_ to do inference and writing results to reused_output_tensors_.
virtual int	NumInputsOfRuntime ()
	Get number of inputs for this model.
virtual int	NumOutputsOfRuntime ()
	Get number of outputs for this model.
virtual TensorInfo	InputInfoOfRuntime (int index)
	Get input information for this model.
virtual TensorInfo	OutputInfoOfRuntime (int index)
	Get output information for this model.
virtual bool	Initialized () const
	Check if the model is initialized successfully.
virtual void	EnableRecordTimeOfRuntime ()
	This is a debug interface, used to record the time of runtime (backend + h2d + d2h) More...
virtual void	DisableRecordTimeOfRuntime ()
	Disable to record the time of runtime, see EnableRecordTimeOfRuntime() for more detail.
virtual std::map< std::string, float >	PrintStatisInfoOfRuntime ()
	Print the statistic information of runtime in the console, see function EnableRecordTimeOfRuntime() for more detail.
virtual bool	EnabledRecordTimeOfRuntime ()
	Check if the EnableRecordTimeOfRuntime() method is enabled.
virtual double	GetProfileTime ()
	Get profile time of Runtime after the profile process is done.
virtual void	ReleaseReusedBuffer ()
	Release reused input/output buffers.

Public Attributes
std::vector< int >	size
	Argument for image preprocessing step, tuple of (width, height), decide the target size after resize, default size = {224, 224}.
std::vector< float >	mean_vals
	Mean parameters for normalize, size should be the the same as channels, default mean_vals = {0.485f, 0.456f, 0.406f}.
std::vector< float >	std_vals
	Std parameters for normalize, size should be the the same as channels, default std_vals = {0.229f, 0.224f, 0.225f}.
Public Attributes inherited from fastdeploy::FastDeployModel
std::vector< Backend >	valid_cpu_backends = {Backend::ORT}
	Model's valid cpu backends. This member defined all the cpu backends have successfully tested for the model.
std::vector< Backend >	valid_gpu_backends = {Backend::ORT}
std::vector< Backend >	valid_ipu_backends = {}
std::vector< Backend >	valid_timvx_backends = {}
std::vector< Backend >	valid_directml_backends = {}
std::vector< Backend >	valid_ascend_backends = {}
std::vector< Backend >	valid_kunlunxin_backends = {}
std::vector< Backend >	valid_rknpu_backends = {}
std::vector< Backend >	valid_sophgonpu_backends = {}

Detailed Description

Torchvision ResNet series model.

Constructor & Destructor Documentation

ResNet()

fastdeploy::vision::classification::ResNet::ResNet	(	const std::string &	model_file,
		const std::string &	params_file = "",
		const RuntimeOption &	custom_option = RuntimeOption(),
		const ModelFormat &	model_format = ModelFormat::ONNX
	)

Set path of model file and the configuration of runtime.

Parameters

[in]	model_file	Path of model file, e.g ./resnet50.onnx
[in]	params_file	Path of parameter file, e.g ppyoloe/model.pdiparams, if the model format is ONNX, this parameter will be ignored
[in]	custom_option	RuntimeOption for inference, the default will use cpu, and choose the backend defined in "valid_cpu_backends"
[in]	model_format	Model format of the loaded model, default is ONNX format

Member Function Documentation

Predict()

bool fastdeploy::vision::classification::ResNet::Predict ( cv::Mat *  im, ClassifyResult *  result, int  topk = 1  )

virtual

Predict for the input "im", the result will be saved in "result".

Parameters

[in]	im	The input image data, comes from cv::imread(), is a 3-D array with layout HWC, BGR format
[in]	result	Saving the inference result.
[in]	topk	The length of return values, e.g., if topk==2, the result will include the 2 most possible class label for input image.

---

The documentation for this class was generated from the following files:

• /fastdeploy/my_work/FastDeploy/fastdeploy/vision/classification/contrib/resnet.h
• /fastdeploy/my_work/FastDeploy/fastdeploy/vision/classification/contrib/resnet.cc