Class Sequential
The Sequential model is a linear stack of layers. Use model to add layers with Add method. Train function to train the layers with dataset. Predict function to invoke prediction against new data.
Inheritance
Namespace: SiaNet
Assembly: SiaNet.dll
Syntax
public class SequentialConstructors
| Improve this Doc View SourceSequential()
Initializes a new instance of the Sequential class.
Declaration
public Sequential()Properties
| Improve this Doc View SourceLayers
Stach of layers for this model
Declaration
public List<BaseLayer> Layers { get; set; }Property Value
| Type | Description |
|---|---|
| System.Collections.Generic.List<BaseLayer> | The layers. |
LearningHistory
Gets or sets the learning history.
Declaration
public History LearningHistory { get; set; }Property Value
| Type | Description |
|---|---|
| History | The learning history. |
Methods
| Improve this Doc View SourceAdd(BaseLayer)
Function to add layer to the sequential model.
Example use:
var model = new Sequential();
model.EpochEnd += Model_EpochEnd;
model.Add(new Dense(100, ActType.ReLU));
model.Add(new Dense(50, ActType.ReLU));
model.Add(new Dense(1, ActType.Sigmoid));Declaration
public void Add(BaseLayer l)Parameters
| Type | Name | Description |
|---|---|---|
| BaseLayer | l | The l. |
Compile(OptimizerType, LossType, MetricType)
Before training a model, you need to configure the learning process, which is done via the compile method. It receives three arguments:
1. An optimizer.This could be the OptimizerType enum or an instance of the Optimizer class.
2. A loss function.This is the objective that the model will try to minimize. It can be the LossType enum identifier of an existing loss function (such as CategoricalCrossentropy or MeanSquaredError), or it can be an instance of the loss class.
3. A metric function.For any classification problem you will want the Accuracy of model. A metric could be the MetricType enum or instance of metric class.
Declaration
public void Compile(OptimizerType optimizer, LossType loss, MetricType metric)Parameters
| Type | Name | Description |
|---|---|---|
| OptimizerType | optimizer | The optimizer type. |
| LossType | loss | The loss type. |
| MetricType | metric | The metric type. |
Compile(BaseOptimizer, LossType, MetricType)
Before training a model, you need to configure the learning process, which is done via the compile method. It receives three arguments:
1. An optimizer.This could be the OptimizerType enum or an instance of the Optimizer class.
2. A loss function.This is the objective that the model will try to minimize. It can be the LossType enum identifier of an existing loss function (such as CategoricalCrossentropy or MeanSquaredError), or it can be an instance of the loss class.
3. A metric function.For any classification problem you will want the Accuracy of model. A metric could be the MetricType enum or instance of metric class.
Declaration
public void Compile(BaseOptimizer optimizer, LossType loss, MetricType metric)Parameters
| Type | Name | Description |
|---|---|---|
| BaseOptimizer | optimizer | The optimizer instance. |
| LossType | loss | The loss type. |
| MetricType | metric | The metric type. |
LoadModel(String)
Loads the model from the saved json to the Sequential model instance.
Declaration
public static Sequential LoadModel(string filePath)Parameters
| Type | Name | Description |
|---|---|---|
| System.String | filePath | The file path. |
Returns
| Type | Description |
|---|---|
| Sequential |
OnBatchEnd(Int32, Int32, Single, Single)
Called when [batch end].
Declaration
protected void OnBatchEnd(int epoch, int batch, float loss, float metric)Parameters
| Type | Name | Description |
|---|---|---|
| System.Int32 | epoch | The epoch. |
| System.Int32 | batch | The batch. |
| System.Single | loss | The loss. |
| System.Single | metric | The metric. |
OnBatchStart(Int32, Int32)
Called when [batch start].
Declaration
protected void OnBatchStart(int epoch, int batch)Parameters
| Type | Name | Description |
|---|---|---|
| System.Int32 | epoch | The epoch. |
| System.Int32 | batch | The batch. |
OnEpochEnd(Int32, Int64, Single, Single, Single, Single, Int64)
Called when [epoch end].
Declaration
protected void OnEpochEnd(int epoch, long samplesSeenPerSec, float loss, float validationLoss, float metric, float validationMetric, long duration)Parameters
| Type | Name | Description |
|---|---|---|
| System.Int32 | epoch | The epoch. |
| System.Int64 | samplesSeenPerSec | The samples seen per sec. |
| System.Single | loss | The loss. |
| System.Single | validationLoss | The validation loss. |
| System.Single | metric | The metric. |
| System.Single | validationMetric | The validation metric. |
| System.Int64 | duration | The duration. |
OnEpochStart(Int32)
Called when [epoch start].
Declaration
protected void OnEpochStart(int epoch)Parameters
| Type | Name | Description |
|---|---|---|
| System.Int32 | epoch | The epoch. |
OnTrainingEnd(History, Int64)
Called when [training end].
Declaration
protected void OnTrainingEnd(History history, long duration)Parameters
| Type | Name | Description |
|---|---|---|
| History | history | The history. |
| System.Int64 | duration | The duration. |
Predict(DataFrame)
Generates output predictions for the input samples.
Declaration
public Tensor Predict(DataFrame x)Parameters
| Type | Name | Description |
|---|---|---|
| DataFrame | x | The input data frame to run prediction. |
Returns
| Type | Description |
|---|---|
| Tensor |
Predict(DataFrame, Int32)
Generates output predictions for the input samples. Computation is done in batches.
Declaration
public Tensor Predict(DataFrame x, int batch_size)Parameters
| Type | Name | Description |
|---|---|---|
| DataFrame | x | The input data frame to run prediction. |
| System.Int32 | batch_size | Size of the batch. |
Returns
| Type | Description |
|---|---|
| Tensor |
SaveModel(String)
Saves the model in json format to the file path.
Declaration
public void SaveModel(string filePath)Parameters
| Type | Name | Description |
|---|---|---|
| System.String | filePath | The file path. |
Train(DataFrameIter, Int32, Int32, DataFrameIter)
Trains the model for a given number of epochs (iterations on a dataset).
Declaration
public void Train(DataFrameIter train, int epochs, int batchSize, DataFrameIter val = null)Parameters
| Type | Name | Description |
|---|---|---|
| DataFrameIter | train | The train dataset which is an instance of DataFrame Iter. |
| System.Int32 | epochs | Integer. Number of epochs to train the model. An epoch is an iteration over the entire x and y data provided. Note that in conjunction with initial_epoch, epochs is to be understood as "final epoch". The model is not trained for a number of iterations given by epochs, but merely until the epoch of index epochs is reached. |
| System.Int32 | batchSize | Integer or None. Number of samples per gradient update. If unspecified, batch_size will default to 32. |
| DataFrameIter | val | The validation set of data to evaluate the model at every epoch. |
Events
| Improve this Doc View SourceBatchEnd
Occurs when [batch end].
Declaration
public event EventHandler<BatchEndEventArgs> BatchEndEvent Type
| Type | Description |
|---|---|
| System.EventHandler<BatchEndEventArgs> |
BatchStart
Occurs when [on batch start].
Declaration
public event EventHandler<BatchStartEventArgs> BatchStartEvent Type
| Type | Description |
|---|---|
| System.EventHandler<BatchStartEventArgs> |
EpochEnd
Occurs when [on epoch end].
Declaration
public event EventHandler<EpochEndEventArgs> EpochEndEvent Type
| Type | Description |
|---|---|
| System.EventHandler<EpochEndEventArgs> |
EpochStart
Occurs when [on epoch start].
Declaration
public event EventHandler<EpochStartEventArgs> EpochStartEvent Type
| Type | Description |
|---|---|
| System.EventHandler<EpochStartEventArgs> |
TrainingEnd
Occurs when [on training end].
Declaration
public event EventHandler<TrainingEndEventArgs> TrainingEndEvent Type
| Type | Description |
|---|---|
| System.EventHandler<TrainingEndEventArgs> |