GenerationSettings class

Controls the sampling behaviour during text generation.

All fields have sensible defaults suitable for conversational use.

const settings = GenerationSettings(
  temperature: 0.8,
  maxTokens: 256,
);

Parameter	Default	Description
temperature	0.7	Randomness — higher = more creative.
topK	40	Keep only the top‑K most probable tokens.
topP	0.9	Nucleus sampling probability threshold.
maxTokens	128	Maximum number of tokens to generate.
repeatPenalty	1.1	Penalise recently generated tokens.

Constructors

GenerationSettings({double temperature = 0.7, int topK = 40, double topP = 0.9, int maxTokens = 128, double repeatPenalty = 1.1})

Creates generation settings. All parameters are optional and fall back to conservative defaults.

const

Properties

hashCode → int

The hash code for this object.

no setter, inherited

maxTokens → int

Maximum number of tokens the model will generate per call.

final

repeatPenalty → double

Penalises tokens that have already appeared, reducing repetition. 1.0 = no penalty.

final

runtimeType → Type

A representation of the runtime type of the object.

no setter, inherited

temperature → double

Controls randomness. 0.0 = deterministic, > 1.0 = very creative.

final

topK → int

Limits sampling to the top‑K most likely next tokens.

final

topP → double

Nucleus sampling: only tokens whose cumulative probability exceeds topP are considered.

final

Methods

noSuchMethod(Invocation invocation) → dynamic

Invoked when a nonexistent method or property is accessed.

inherited

toMap() → Map<String, dynamic>

Serialises the settings to a Map for the platform channel.

toString() → String

A string representation of this object.

inherited

Operators

operator ==(Object other) → bool

The equality operator.

inherited