Provides methods for neural network model-based training in the batch processing mode.
More...
template<typename algorithmFPType, Method method>
class daal::algorithms::neural_networks::training::interface1::Distributed< step1Local, algorithmFPType, method >
- Template Parameters
-
algorithmFPType | Data type to use in intermediate computations for neural network model-based training, double or float |
method | Neural network training method, training::Method |
- Enumerations
-
- References
-
◆ Distributed() [1/2]
◆ Distributed() [2/2]
Constructs neural network by copying input objects and parameters of another neural network
- Parameters
-
[in] | other | An algorithm to be used as the source to initialize the input objects and parameters of the algorithm |
◆ clone()
Returns a pointer to the newly allocated neural network with a copy of input objects and parameters of this neural network
- Returns
- Pointer to the newly allocated layer
◆ getMethod()
virtual int getMethod |
( |
| ) |
const |
|
inlinevirtual |
Returns method of the algorithm
- Returns
- Method of the algorithm
Implements AlgorithmIface.
◆ getPartialResult()
Returns structure that contains computed partial results of the neural network algorithm
- Returns
- Structure that contains partial results of the neural network algorithm
◆ getResult()
Returns the structure that contains the results of the neural network algorithm
- Returns
- Structure that contains the results of the neural network algorithm
◆ setPartialResult()
Registers user-allocated memory to store partial results of the neural network algorithm
- Parameters
-
[in] | partialResult | Structure for storing partial results of the neural network algorithm |
- Returns
- Status of computations
◆ setResult()
Register user-allocated memory to store the results of the neural network algorithm
- Parameters
-
[in] | res | Structure for storing results of the neural network algorithm |
- Returns
- Status of computations
◆ input
◆ parameter
The documentation for this class was generated from the following file: