QATrainer.Options Class

Definition

Namespace:

Microsoft.ML.TorchSharp.Roberta

Assembly:

Microsoft.ML.TorchSharp.dll

Package:

Microsoft.ML.TorchSharp v0.21.1

public sealed class QATrainer.Options : Microsoft.ML.TorchSharp.NasBert.NasBertTrainer.NasBertOptions

type QATrainer.Options = class
    inherit NasBertTrainer.NasBertOptions

Public NotInheritable Class QATrainer.Options
Inherits NasBertTrainer.NasBertOptions

Inheritance

Object

TransformInputBase

TorchSharpBaseTrainer.Options

NasBertTrainer.NasBertOptions

QATrainer.Options

Constructors

QATrainer.Options()

Fields

ActivationDropout	Dropout rate after activation functions in FFN layers. Should be within [0, 1). (Inherited from NasBertTrainer.NasBertOptions)
AdamBetas	Betas for Adam optimizer. (Inherited from NasBertTrainer.NasBertOptions)
AdamEps	Epsilon for Adam optimizer. (Inherited from NasBertTrainer.NasBertOptions)
AnswerIndexStartColumnName	Answer Index Start Column Name
AttentionDropout	Dropout rate for attention weights. Should be within [0, 1). (Inherited from NasBertTrainer.NasBertOptions)
BatchSize	Number of samples to use for mini-batch training. (Inherited from TorchSharpBaseTrainer.Options)
ClipNorm	The clipping threshold of gradients. Should be within [0, +Inf). 0 means not to clip norm. (Inherited from NasBertTrainer.NasBertOptions)
ContextColumnName	Context Column Name
Dropout	Dropout rate for general situations. Should be within [0, 1). (Inherited from NasBertTrainer.NasBertOptions)
DynamicDropout	Whether to use dynamic dropout. (Inherited from NasBertTrainer.NasBertOptions)
EncoderNormalizeBefore	Whether to apply layer normalization before each encoder block. (Inherited from NasBertTrainer.NasBertOptions)
FinalLearningRateRatio	The final learning rate for polynomial decay scheduler. (Inherited from TorchSharpBaseTrainer.Options)
FreezeEncoder	Whether to freeze encoder parameters. (Inherited from NasBertTrainer.NasBertOptions)
FreezeTransfer	Whether to freeze transfer module parameters. (Inherited from NasBertTrainer.NasBertOptions)
LabelColumnName	The label column name. (Inherited from TorchSharpBaseTrainer.Options)
LayerNormTraining	Whether to train layer norm parameters. (Inherited from NasBertTrainer.NasBertOptions)
LearningRate	Learning rate for the first N epochs; all epochs >N using LR_N. Note: this may be interpreted differently depending on the scheduler. (Inherited from NasBertTrainer.NasBertOptions)
LogEveryNStep	How often to log the loss.
MaxEpoch	Stop training when reaching this number of epochs. (Inherited from TorchSharpBaseTrainer.Options)
PoolerDropout	Dropout rate in the masked language model pooler layers. Should be within [0, 1). (Inherited from NasBertTrainer.NasBertOptions)
PredictedAnswerColumnName	Answer Column Name for the predicted answers
PredictionColumnName	The Prediction column name. (Inherited from TorchSharpBaseTrainer.Options)
QuestionColumnName	Question Column Name
ScoreColumnName	The Score column name. (Inherited from TorchSharpBaseTrainer.Options)
Sentence1ColumnName	The first sentence column. (Inherited from NasBertTrainer.NasBertOptions)
Sentence2ColumnName	The second sentence column. (Inherited from NasBertTrainer.NasBertOptions)
StartLearningRateRatio	The start learning rate for polynomial decay scheduler. (Inherited from TorchSharpBaseTrainer.Options)
TaskType	Task type, which is related to the model head. (Inherited from NasBertTrainer.NasBertOptions)
TopKAnswers	Number of top predicted answers in question answering task.
TrainingAnswerColumnName	Answer Column Name for the training data
ValidationSet	The validation set used while training to improve model quality. (Inherited from TorchSharpBaseTrainer.Options)
WarmupRatio	Proportion of warmup steps for polynomial decay scheduler. (Inherited from NasBertTrainer.NasBertOptions)
WeightDecay	Coefficiency of weight decay. Should be within [0, +Inf). (Inherited from TorchSharpBaseTrainer.Options)