AnalysisTemplate.AnalysisTemplate Class Reference

Collaboration diagram for AnalysisTemplate.AnalysisTemplate:

Public Member Functions
	StartAnalysis (self, str args)
	EndAnalysis (self)
	StartNetworkExecution (self, list inputs)
	EndNetworkExecution (self, list outputs)
	DefaultOpPre (self, str name, str opcode, list inputs)
	DefaultOpPost (self, str name, str opcode, list inputs, dict output)
	Conv2DPre (self, str name, dict input, dict filter, dict bias, str padding, dict stride, dict dilation, str fused_act)
	Conv2DPost (self, str name, dict input, dict filter, dict bias, str padding, dict stride, dict dilation, dict output, str fused_act)

Detailed Description

Definition at line 25 of file AnalysisTemplate.py.

Member Function Documentation

Conv2DPost()

AnalysisTemplate.AnalysisTemplate.Conv2DPost	(		self,
		str	name,
		dict	input,
		dict	filter,
		dict	bias,
		str	padding,
		dict	stride,
		dict	dilation,
		dict	output,
		str	fused_act
	)

Called after Conv2D layer execution
name: output tensor name (string)
opcode: opcode name (string)
input: Tensor
filter: Tensor
bias: Tensor
padding: Padding (string)
stride: Stride
dilation: Dilation
output: Tensor
fused_act: Fused activation functions (string)

Definition at line 101 of file AnalysisTemplate.py.

                   stride: dict, dilation: dict, output: dict, fused_act: str):
        """
        Called after Conv2D layer execution
        name: output tensor name (string)
        opcode: opcode name (string)
        input: Tensor
        filter: Tensor
        bias: Tensor
        padding: Padding (string)
        stride: Stride
        dilation: Dilation
        output: Tensor
        fused_act: Fused activation functions (string)
        """
        print("name", name)
        print("input", input)
        print("filter", filter)
        print("bias", bias)
        print("padding", padding)
        print("stride", stride)
        print("dilation", dilation)
        print("output shape", output['data'].shape)
        print("output type", output['data'].dtype)
        print("fused activation", fused_act)

Conv2DPre()

AnalysisTemplate.AnalysisTemplate.Conv2DPre	(		self,
		str	name,
		dict	input,
		dict	filter,
		dict	bias,
		str	padding,
		dict	stride,
		dict	dilation,
		str	fused_act
	)

Called before Conv2D layer execution
name: output tensor name (string)
opcode: opcode name (string)
input: Tensor
filter: Tensor
bias: Tensor
padding: Padding (string)
stride: Stride
dilation: Dilation
fused_act: Fused activation functions (string)

Definition at line 78 of file AnalysisTemplate.py.

                  stride: dict, dilation: dict, fused_act: str):
        """
        Called before Conv2D layer execution
        name: output tensor name (string)
        opcode: opcode name (string)
        input: Tensor
        filter: Tensor
        bias: Tensor
        padding: Padding (string)
        stride: Stride
        dilation: Dilation
        fused_act: Fused activation functions (string)
        """
        print("name", name)
        print("input", input)
        print("filter", filter)
        print("bias", bias)
        print("padding", padding)
        print("stride", stride)
        print("dilation", dilation)
        print("fused activation", fused_act)
 

DefaultOpPost()

AnalysisTemplate.AnalysisTemplate.DefaultOpPost	(		self,
		str	name,
		str	opcode,
		list	inputs,
		dict	output
	)

Default hook called after an operator is executed
name: output tensor name (string)
opcode: opcode name (string)
inputs: list of Tensor
output: Tensor

Definition at line 65 of file AnalysisTemplate.py.

    def DefaultOpPost(self, name: str, opcode: str, inputs: list, output: dict):
        """
        Default hook called after an operator is executed
        name: output tensor name (string)
        opcode: opcode name (string)
        inputs: list of Tensor
        output: Tensor
        """
        print("name", name)
        print("opcode", opcode)
        print("inputs", inputs)
        print("output", output)
 

DefaultOpPre()

AnalysisTemplate.AnalysisTemplate.DefaultOpPre	(		self,
		str	name,
		str	opcode,
		list	inputs
	)

Default hook called before an operator is executed
name: output tensor name (string)
opcode: opcode name (string)
inputs: list of Tensor

Definition at line 54 of file AnalysisTemplate.py.

    def DefaultOpPre(self, name: str, opcode: str, inputs: list):
        """
        Default hook called before an operator is executed
        name: output tensor name (string)
        opcode: opcode name (string)
        inputs: list of Tensor
        """
        print("name", name)
        print("opcode", opcode)
        print("inputs", inputs)
 

EndAnalysis()

AnalysisTemplate.AnalysisTemplate.EndAnalysis

(

self

)

Called when the analysis ends

Definition at line 34 of file AnalysisTemplate.py.

    def EndAnalysis(self):
        """
        Called when the analysis ends
        """
        print("Analysis ended.")
 

EndNetworkExecution()

AnalysisTemplate.AnalysisTemplate.EndNetworkExecution	(		self,
		list	outputs
	)

Called when the execution of a network ends
outputs: list of Tensor

Definition at line 47 of file AnalysisTemplate.py.

    def EndNetworkExecution(self, outputs: list):
        """
        Called when the execution of a network ends
        outputs: list of Tensor
        """
        print("Network execution ended.")
 

StartAnalysis()

AnalysisTemplate.AnalysisTemplate.StartAnalysis	(		self,
		str	args
	)

Called when the analysis starts
args: string given by --analysis_args option

Definition at line 26 of file AnalysisTemplate.py.

    def StartAnalysis(self, args: str):
        """
        Called when the analysis starts
        args: string given by --analysis_args option
        """
        print("Analysis started.")
        print("args", args)
 

StartNetworkExecution()

AnalysisTemplate.AnalysisTemplate.StartNetworkExecution	(		self,
		list	inputs
	)

Called when the execution of a network starts
inputs: list of Tensor

Definition at line 40 of file AnalysisTemplate.py.

    def StartNetworkExecution(self, inputs: list):
        """
        Called when the execution of a network starts
        inputs: list of Tensor
        """
        print("Network execution started.")
 

---

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

• compiler/dalgona/analysis/AnalysisTemplate.py