Struct FTryllSamplingOverrides

ClassList > FTryllSamplingOverrides

More...

• #include <TryllSamplingOverrides.h>

Public Attributes

Type	Name
float	FrequencyPenalty = 0.0
int32	MaxTokens = 0
float	MinP = 0.0
float	PresencePenalty = 0.0
float	RepeatPenalty = 0.0
int32	Seed = 0
float	Temperature = 0.0
int32	TopK = 0
float	TopP = 0.0
bool	bOverrideFrequencyPenalty = false
bool	bOverrideMaxTokens = false
bool	bOverrideMinP = false
bool	bOverridePresencePenalty = false
bool	bOverrideRepeatPenalty = false
bool	bOverrideSeed = false
bool	bOverrideTemperature = false
bool	bOverrideTopK = false
bool	bOverrideTopP = false

Detailed Description

Sparse sampling override block. Null on any field means "inherit the selected model's default". No per-field (on_change) attributes: the outer sub-table field carries one callback that re-resolves the whole block and commits derived state.

Public Attributes Documentation

variable FrequencyPenalty

float FTryllSamplingOverrides::FrequencyPenalty;

---

variable MaxTokens

int32 FTryllSamplingOverrides::MaxTokens;

---

variable MinP

float FTryllSamplingOverrides::MinP;

---

variable PresencePenalty

float FTryllSamplingOverrides::PresencePenalty;

---

variable RepeatPenalty

float FTryllSamplingOverrides::RepeatPenalty;

---

variable Seed

int32 FTryllSamplingOverrides::Seed;

---

variable Temperature

float FTryllSamplingOverrides::Temperature;

---

variable TopK

int32 FTryllSamplingOverrides::TopK;

---

variable TopP

float FTryllSamplingOverrides::TopP;

---

variable bOverrideFrequencyPenalty

bool FTryllSamplingOverrides::bOverrideFrequencyPenalty;

OpenAI-style frequency penalty. Null = inherit model default.

---

variable bOverrideMaxTokens

bool FTryllSamplingOverrides::bOverrideMaxTokens;

Maximum new tokens to generate. Null = inherit model default.

---

variable bOverrideMinP

bool FTryllSamplingOverrides::bOverrideMinP;

Minimum token probability (min-p sampling). Null = inherit model default.

---

variable bOverridePresencePenalty

bool FTryllSamplingOverrides::bOverridePresencePenalty;

OpenAI-style presence penalty. Null = inherit model default.

---

variable bOverrideRepeatPenalty

bool FTryllSamplingOverrides::bOverrideRepeatPenalty;

Penalise repeated tokens (llama.cpp repeat_penalty). Null = inherit.

---

variable bOverrideSeed

bool FTryllSamplingOverrides::bOverrideSeed;

RNG seed for reproducible sampling. Null = inherit model default.

---

variable bOverrideTemperature

bool FTryllSamplingOverrides::bOverrideTemperature;

0.0 = fully deterministic, 2.0 = highly random. Null = inherit.

---

variable bOverrideTopK

bool FTryllSamplingOverrides::bOverrideTopK;

Top-k sampling limit. Null = inherit model default.

---

variable bOverrideTopP

bool FTryllSamplingOverrides::bOverrideTopP;

Nucleus sampling cutoff. Null = inherit model default.

---

---

The documentation for this class was generated from the following file C:/_tryll/_monorepo3/tryll/clients/unreal/Source/TryllClient/Public/Generated/Nodes/TryllSamplingOverrides.h