Type alias: LlamaChatSessionRepeatPenalty ​

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

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

Type declaration ​

frequencyPenalty ​

frequencyPenalty?: number;

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.

lastTokens ​

lastTokens?: number;

lastTokens?: number;

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

penalizeNewLine ​

penalizeNewLine?: boolean;

penalizeNewLine?: boolean;

Penalize new line tokens. Enabled by default.

penalty ​

penalty?: number;

penalty?: number;

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

presencePenalty ​

presencePenalty?: number;

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.

punishTokensFilter ​

punishTokensFilter?: (tokens) => Token[];

punishTokensFilter?: (tokens) => Token[];

Parameters ​

Source ​

llamaEvaluator/LlamaChatSession.ts:79