Skip to content

Type Alias: LlamaChatSessionRepeatPenalty

ts
type LlamaChatSessionRepeatPenalty: {
  lastTokens: number;
  punishTokensFilter: (tokens: Token[]) => Token[];
  penalizeNewLine: boolean;
  penalty: number;
  frequencyPenalty: number;
  presencePenalty: number;
};

Type declaration

lastTokens?

ts
optional lastTokens: number;

Number of recent tokens generated by the model to apply penalties to repetition of. Defaults to 64.

punishTokensFilter()?

ts
optional punishTokensFilter: (tokens: Token[]) => Token[];

Parameters

ParameterType
tokensToken[]

Returns

Token[]

penalizeNewLine?

ts
optional penalizeNewLine: boolean;

Penalize new line tokens. Enabled by default.

penalty?

ts
optional penalty: number;

The relative amount to lower the probability of the tokens in punishTokens by Defaults to 1.1. Set to 1 to disable.

frequencyPenalty?

ts
optional frequencyPenalty: number;

For n time a token is in the punishTokens array, lower its probability by n * frequencyPenalty Disabled by default (0). Set to a value between 0 and 1 to enable.

presencePenalty?

ts
optional presencePenalty: number;

Lower the probability of all the tokens in the punishTokens array by presencePenalty Disabled by default (0). Set to a value between 0 and 1 to enable.

Defined in

evaluator/LlamaChatSession/LlamaChatSession.ts:251