Class: GgufInsightsConfigurationResolver

Defined in: gguf/insights/GgufInsightsConfigurationResolver.ts:16

Accessors

ggufInsights

Get Signature

get ggufInsights(): GgufInsights;

Defined in: gguf/insights/GgufInsightsConfigurationResolver.ts:23

Returns

GgufInsights

Methods

resolveAndScoreConfig()

resolveAndScoreConfig(options: {
  targetGpuLayers?: number | "max";
  targetContextSize?: number;
  embeddingContext?: boolean;
  flashAttention?: boolean;
  swaFullCache?: boolean;
  useMmap?: boolean;
}, hardwareOverrides: {
  getVramState?: Promise<{
     total: number;
     free: number;
     unifiedSize: number;
  }>;
  getRamState?: Promise<{
     total: number;
     free: number;
  }>;
  getSwapState?: Promise<{
     total: number;
     free: number;
  }>;
  llamaVramPaddingSize?: number;
  llamaGpu?: false | "metal" | "cuda" | "vulkan";
  llamaSupportsGpuOffloading?: boolean;
}): Promise<{
  compatibilityScore: number;
  bonusScore: number;
  totalScore: number;
  resolvedValues: {
     gpuLayers: number;
     contextSize: number;
     modelRamUsage: number;
     contextRamUsage: number;
     totalRamUsage: number;
     modelVramUsage: number;
     contextVramUsage: number;
     totalVramUsage: number;
  };
}>;

Defined in: gguf/insights/GgufInsightsConfigurationResolver.ts:37

Resolve the best configuration for loading a model and creating a context using the current hardware.

Specifying a targetGpuLayers and/or targetContextSize will ensure the resolved configuration matches those values, but note it can lower the compatibility score if the hardware doesn't support it.

Overriding hardware values it possible by configuring hardwareOverrides.

Parameters

Parameter	Type	Description
options	{ targetGpuLayers?: number | "max"; targetContextSize?: number; embeddingContext?: boolean; flashAttention?: boolean; swaFullCache?: boolean; useMmap?: boolean; }
options.targetGpuLayers?	number | "max"	-
options.targetContextSize?	number	-
options.embeddingContext?	boolean	-
options.flashAttention?	boolean	-
options.swaFullCache?	boolean	-
options.useMmap?	boolean	-
hardwareOverrides	{ getVramState?: Promise<{ total: number; free: number; unifiedSize: number; }>; getRamState?: Promise<{ total: number; free: number; }>; getSwapState?: Promise<{ total: number; free: number; }>; llamaVramPaddingSize?: number; llamaGpu?: false | "metal" | "cuda" | "vulkan"; llamaSupportsGpuOffloading?: boolean; }
hardwareOverrides.getVramState?	-
hardwareOverrides.getRamState?	-
hardwareOverrides.getSwapState?	-
hardwareOverrides.llamaVramPaddingSize?	number	-
hardwareOverrides.llamaGpu?	false | "metal" | "cuda" | "vulkan"	-
hardwareOverrides.llamaSupportsGpuOffloading?	boolean	-

Returns

Promise<{ compatibilityScore: number; bonusScore: number; totalScore: number; resolvedValues: { gpuLayers: number; contextSize: number; modelRamUsage: number; contextRamUsage: number; totalRamUsage: number; modelVramUsage: number; contextVramUsage: number; totalVramUsage: number; }; }>

---

scoreModelConfigurationCompatibility()

scoreModelConfigurationCompatibility(__namedParameters: {
  contextSize?: number;
  embeddingContext?: boolean;
  flashAttention?: boolean;
  swaFullCache?: boolean;
  maximumFittedContextSizeMultiplier?: number;
  maximumUnfitConfigurationResourceMultiplier?: number;
  forceStrictContextSize?: boolean;
  forceGpuLayers?: number | "max";
  useMmap?: boolean;
}, __namedParameters: {
  getVramState?: Promise<{
     total: number;
     free: number;
     unifiedSize: number;
  }>;
  getRamState?: Promise<{
     total: number;
     free: number;
  }>;
  getSwapState?: Promise<{
     total: number;
     free: number;
  }>;
  llamaVramPaddingSize?: number;
  llamaGpu?: false | "metal" | "cuda" | "vulkan";
  llamaSupportsGpuOffloading?: boolean;
}): Promise<{
  compatibilityScore: number;
  bonusScore: number;
  totalScore: number;
  resolvedValues: {
     gpuLayers: number;
     contextSize: number;
     modelRamUsage: number;
     contextRamUsage: number;
     totalRamUsage: number;
     modelVramUsage: number;
     contextVramUsage: number;
     totalVramUsage: number;
  };
}>;

Defined in: gguf/insights/GgufInsightsConfigurationResolver.ts:107

Score the compatibility of the model configuration with the current GPU and VRAM state. Assumes a model is loaded with the default "auto" configurations. Scored based on the following criteria:

• The number of GPU layers that can be offloaded to the GPU (only if there's a GPU. If there's no GPU then by how small the model is)
• Whether all layers can be offloaded to the GPU (gives additional points)
• Whether the resolved context size is at least as large as the specified contextSize

If the resolved context size is larger than the specified context size, for each multiplier of the specified contextSize that the resolved context size is larger by, 1 bonus point is given in the bonusScore.

maximumFittedContextSizeMultiplier is used to improve the proportionality of the bonus score between models. Set this to any value higher than <max compared model context size> / contextSize. Defaults to 100.

maximumUnfitConfigurationResourceMultiplier is used to improve the proportionality of the bonus score between unfit models. Set this to any value higher than <max compared model resource usage> / <total available resources>. Defaults to 100.

contextSize defaults to 4096 (if the model train context size is lower than this, the model train context size is used instead).

Parameters

Parameter	Type	Description
__namedParameters	{ contextSize?: number; embeddingContext?: boolean; flashAttention?: boolean; swaFullCache?: boolean; maximumFittedContextSizeMultiplier?: number; maximumUnfitConfigurationResourceMultiplier?: number; forceStrictContextSize?: boolean; forceGpuLayers?: number | "max"; useMmap?: boolean; }	-
__namedParameters.contextSize?	number	-
__namedParameters.embeddingContext?	boolean	-
__namedParameters.flashAttention?	boolean	-
__namedParameters.swaFullCache?	boolean	-
__namedParameters.maximumFittedContextSizeMultiplier?	number	-
__namedParameters.maximumUnfitConfigurationResourceMultiplier?	number	-
__namedParameters.forceStrictContextSize?	boolean	Do not resolve a context size larger than the specified contextSize. Defaults to false.
__namedParameters.forceGpuLayers?	number | "max"	-
__namedParameters.useMmap?	boolean	-
__namedParameters	{ getVramState?: Promise<{ total: number; free: number; unifiedSize: number; }>; getRamState?: Promise<{ total: number; free: number; }>; getSwapState?: Promise<{ total: number; free: number; }>; llamaVramPaddingSize?: number; llamaGpu?: false | "metal" | "cuda" | "vulkan"; llamaSupportsGpuOffloading?: boolean; }	-
__namedParameters.getVramState?	-
__namedParameters.getRamState?	-
__namedParameters.getSwapState?	-
__namedParameters.llamaVramPaddingSize?	number	-
__namedParameters.llamaGpu?	false | "metal" | "cuda" | "vulkan"	-
__namedParameters.llamaSupportsGpuOffloading?	boolean	-

Returns

Promise<{ compatibilityScore: number; bonusScore: number; totalScore: number; resolvedValues: { gpuLayers: number; contextSize: number; modelRamUsage: number; contextRamUsage: number; totalRamUsage: number; modelVramUsage: number; contextVramUsage: number; totalVramUsage: number; }; }>

---

resolveModelGpuLayers()

resolveModelGpuLayers(gpuLayers?: 
  | number
  | "auto"
  | "max"
  | {
  min?: number;
  max?: number;
  fitContext?: {
     contextSize?: number;
     embeddingContext?: boolean;
  };
}, __namedParameters?: {
  ignoreMemorySafetyChecks?: boolean;
  getVramState?: Promise<{
     total: number;
     free: number;
  }>;
  llamaVramPaddingSize?: number;
  llamaGpu?: false | "metal" | "cuda" | "vulkan";
  llamaSupportsGpuOffloading?: boolean;
  defaultContextFlashAttention?: boolean;
  defaultContextSwaFullCache?: boolean;
  useMmap?: boolean;
}): Promise<number>;

Defined in: gguf/insights/GgufInsightsConfigurationResolver.ts:376

Parameters

Parameter	Type
gpuLayers?	| number | "auto" | "max" | { min?: number; max?: number; fitContext?: { contextSize?: number; embeddingContext?: boolean; }; }
__namedParameters?	{ ignoreMemorySafetyChecks?: boolean; getVramState?: Promise<{ total: number; free: number; }>; llamaVramPaddingSize?: number; llamaGpu?: false | "metal" | "cuda" | "vulkan"; llamaSupportsGpuOffloading?: boolean; defaultContextFlashAttention?: boolean; defaultContextSwaFullCache?: boolean; useMmap?: boolean; }
__namedParameters.ignoreMemorySafetyChecks?	boolean
__namedParameters.getVramState?
__namedParameters.llamaVramPaddingSize?	number
__namedParameters.llamaGpu?	false | "metal" | "cuda" | "vulkan"
__namedParameters.llamaSupportsGpuOffloading?	boolean
__namedParameters.defaultContextFlashAttention?	boolean
__namedParameters.defaultContextSwaFullCache?	boolean
__namedParameters.useMmap?	boolean

Returns

Promise<number>

---

resolveContextContextSize()

resolveContextContextSize(contextSize: 
  | undefined
  | number
  | "auto"
  | {
  min?: number;
  max?: number;
}, __namedParameters: {
  modelGpuLayers: number;
  modelTrainContextSize: number;
  flashAttention?: boolean;
  swaFullCache?: boolean;
  batchSize?: number;
  sequences?: number;
  getVramState?: Promise<{
     total: number;
     free: number;
     unifiedSize: number;
  }>;
  getRamState?: Promise<{
     total: number;
     free: number;
  }>;
  getSwapState?: Promise<{
     total: number;
     free: number;
  }>;
  llamaGpu?: false | "metal" | "cuda" | "vulkan";
  ignoreMemorySafetyChecks?: boolean;
  isEmbeddingContext?: boolean;
}): Promise<number>;

Defined in: gguf/insights/GgufInsightsConfigurationResolver.ts:407

Resolve a context size option for the given options and constraints.

If there's no context size that can fit the available resources, an InsufficientMemoryError is thrown.

Parameters

Parameter	Type
contextSize	| undefined | number | "auto" | { min?: number; max?: number; }
__namedParameters	{ modelGpuLayers: number; modelTrainContextSize: number; flashAttention?: boolean; swaFullCache?: boolean; batchSize?: number; sequences?: number; getVramState?: Promise<{ total: number; free: number; unifiedSize: number; }>; getRamState?: Promise<{ total: number; free: number; }>; getSwapState?: Promise<{ total: number; free: number; }>; llamaGpu?: false | "metal" | "cuda" | "vulkan"; ignoreMemorySafetyChecks?: boolean; isEmbeddingContext?: boolean; }
__namedParameters.modelGpuLayers	number
__namedParameters.modelTrainContextSize	number
__namedParameters.flashAttention?	boolean
__namedParameters.swaFullCache?	boolean
__namedParameters.batchSize?	number
__namedParameters.sequences?	number
__namedParameters.getVramState?
__namedParameters.getRamState?
__namedParameters.getSwapState?
__namedParameters.llamaGpu?	false | "metal" | "cuda" | "vulkan"
__namedParameters.ignoreMemorySafetyChecks?	boolean
__namedParameters.isEmbeddingContext?	boolean

Returns

Promise<number>