Documentation
¶
Overview ¶
Package layers contains shared things between layers. It also contains functions that will be supplimental to cudnn.
Index ¶
- type Concat
- type Tensor
- func BuildRandomTensor(handle *cudnn.Handler, frmt gocudnn.TensorFormat, dtype gocudnn.DataType, ...) (*Tensor, error)
- func CreateTensor(handle *cudnn.Handler, frmt gocudnn.TensorFormat, dtype gocudnn.DataType, ...) (t *Tensor, err error)
- func ZeroClone(h *cudnn.Handler, t *Tensor) (c *Tensor, err error)
- func (t *Tensor) AvgX(handle *cudnn.Handler) (x float32, err error)
- func (t *Tensor) LoadValuesFromSLice(handle *cudnn.Handler, input interface{}, length int32) error
- func (t *Tensor) MaxX(handle *cudnn.Handler) (x float32, err error)
- func (t *Tensor) MinX(handle *cudnn.Handler) (x float32, err error)
- func (t *Tensor) Norm1X(handle *cudnn.Handler) (x float32, err error)
- func (t *Tensor) Norm2X(handle *cudnn.Handler) (x float32, err error)
Constants ¶
This section is empty.
Variables ¶
This section is empty.
Functions ¶
This section is empty.
Types ¶
type Concat ¶
type Concat struct {
// contains filtered or unexported fields
}
Concat does the concat operation
func CreateConcat ¶
CreateConcat creates a concat operation handler
func (*Concat) FindOutputDims ¶
FindOutputDims finds the output dims
type Tensor ¶
Tensor is a tensor with volume
func BuildRandomTensor ¶
func BuildRandomTensor(handle *cudnn.Handler, frmt gocudnn.TensorFormat, dtype gocudnn.DataType, dims []int32, mean, std float32) (*Tensor, error)
BuildRandomTensor builds a fix sized input
func CreateTensor ¶
func CreateTensor(handle *cudnn.Handler, frmt gocudnn.TensorFormat, dtype gocudnn.DataType, dims []int32) (t *Tensor, err error)
CreateTensor creates a tensor.
func (*Tensor) LoadValuesFromSLice ¶
LoadValuesFromSLice takes a go slice and fills it into the tensor sitting in the gpu. If the length of goslice doesn't fit the input it will return an error
func (*Tensor) MaxX ¶
MaxX returns the MaxX per batch value in the tensor or if it is used for the filter it would be the MaxX value per neuron
func (*Tensor) MinX ¶
MinX returns the minx value per batch in the tensor or if it is used for the filter it would be the minx value per neuron
Directories
¶
Path | Synopsis |
---|---|
Package cnn contains structs and methods used to do forward, and backward operations for convolution layers
|
Package cnn contains structs and methods used to do forward, and backward operations for convolution layers |