com.volcengine.model.maas.api

Interface Api.ParametersOrBuilder

All Superinterfaces:

com.google.protobuf.MessageLiteOrBuilder, com.google.protobuf.MessageOrBuilder

All Known Implementing Classes:

Api.Parameters, Api.Parameters.Builder

Enclosing class:

Api

public static interface Api.ParametersOrBuilder
extends com.google.protobuf.MessageOrBuilder

Method Summary

Modifier and Type	Method and Description
boolean	getDoSample() Whether or not to use sampling, use greedy decoding otherwise.
float	getFrequencyPenalty() Number between -2.0 and 2.0.
long	getMaxNewTokens() The maximum number of tokens to generate, ignoring the number of tokens in the prompt
long	getMaxTokens() The maximum number of tokens to generate in the char completion.
float	getPresencePenalty() Number between -2.0 and 2.0.
float	getRepetitionPenalty() The parameter for repetition penalty, from [1.0, 2.0]
float	getTemperature() Exponential scaling output probability distribution
long	getTopK() The number of highest probability vocabulary tokens to keep for top-k-filtering.
float	getTopP() An alternative to sampling with temperature, called nucleus sampling, where the model considers the results of the tokens with top_p probability mass

Methods inherited from interface com.google.protobuf.MessageOrBuilder

findInitializationErrors, getAllFields, getDefaultInstanceForType, getDescriptorForType, getField, getInitializationErrorString, getOneofFieldDescriptor, getRepeatedField, getRepeatedFieldCount, getUnknownFields, hasField, hasOneof

Methods inherited from interface com.google.protobuf.MessageLiteOrBuilder

isInitialized

Method Detail

getTemperature

float getTemperature()

 Exponential scaling output probability distribution
 

float temperature = 1;

Returns:

The temperature.

getMaxTokens

long getMaxTokens()

 The maximum number of tokens to generate in the char completion.
 

int64 max_tokens = 2;

Returns:

The maxTokens.

getTopP

float getTopP()

 An alternative to sampling with temperature, called nucleus sampling,
 where the model considers the results of the tokens with top_p probability
 mass
 

float top_p = 3;

Returns:

The topP.

getPresencePenalty

float getPresencePenalty()

 Number between -2.0 and 2.0. Positive values penalize new tokens based
 on whether they appear in the text so far, increasing the model's
 likelihood to talk about new topics.
 

float presence_penalty = 4;

Returns:

The presencePenalty.

getFrequencyPenalty

float getFrequencyPenalty()

 Number between -2.0 and 2.0. Positive values penalize new tokens based on
 their existing frequency in the text so far, decreasing the model's
 likelihood to repeat the same line verbatim.
 

float frequency_penalty = 5;

Returns:

The frequencyPenalty.

getMaxNewTokens

long getMaxNewTokens()

 The maximum number of tokens to generate, ignoring the number of tokens in
 the prompt
 

int64 max_new_tokens = 6;

Returns:

The maxNewTokens.

getRepetitionPenalty

float getRepetitionPenalty()

 The parameter for repetition penalty, from [1.0, 2.0]
 

float repetition_penalty = 7;

Returns:

The repetitionPenalty.

getDoSample

boolean getDoSample()

 Whether or not to use sampling, use greedy decoding otherwise. Default to
 false
 

bool do_sample = 8;

Returns:

The doSample.

getTopK

long getTopK()

 The number of highest probability vocabulary tokens to keep for
 top-k-filtering.
 

int64 top_k = 9;

Returns:

The topK.