node-llama-cpp 2.8.6 → 3.0.0-beta.10

package/dist/llamaEvaluator/LlamaContext.js

@@ -1,143 +0,0 @@
-import { removeNullFields } from "../utils/removeNullFields.js";
-import { LLAMAContext } from "./LlamaBins.js";
-export class LlamaContext {
-    _model;
-    _ctx;
-    _prependBos;
-    _prependTokens;
-    /** @internal */
-    _chatGrammar;
-    /**
-     * @param {LlamaContextOptions} options
-     */
-    constructor({ model, prependBos = true, grammar, seed = model._contextOptions.seed, contextSize = model._contextOptions.contextSize, batchSize = model._contextOptions.batchSize, logitsAll = model._contextOptions.logitsAll, embedding = model._contextOptions.embedding, threads = model._contextOptions.threads }) {
-        this._model = model;
-        this._ctx = new LLAMAContext(model._model, removeNullFields({
-            seed: seed != null ? Math.max(-1, seed) : undefined,
-            contextSize,
-            batchSize,
-            logitsAll,
-            embedding,
-            threads
-        }));
-        this._prependBos = prependBos;
-        this._prependTokens = [];
-        this._chatGrammar = grammar;
-        if (prependBos) {
-            this._prependTokens.unshift(this._ctx.tokenBos());
-        }
-    }
-    encode(text) {
-        if (text === "")
-            return new Uint32Array();
-        return this._ctx.encode(text);
-    }
-    decode(tokens) {
-        if (tokens.length === 0)
-            return "";
-        if (tokens instanceof Uint32Array)
-            return this._ctx.decode(tokens);
-        return this._ctx.decode(Uint32Array.from(tokens));
-    }
-    get prependBos() {
-        return this._prependBos;
-    }
-    /**
-     * @returns {Token | null} The BOS (Beginning Of Sequence) token.
-     */
-    getBosToken() {
-        const bosToken = this._ctx.tokenBos();
-        if (bosToken === -1)
-            return null;
-        return bosToken;
-    }
-    /**
-     * @returns {Token | null} The EOS (End Of Sequence) token.
-     */
-    getEosToken() {
-        const eosToken = this._ctx.tokenEos();
-        if (eosToken === -1)
-            return null;
-        return eosToken;
-    }
-    /**
-     * @returns {Token | null} The NL (New Line) token.
-     */
-    getNlToken() {
-        const nlToken = this._ctx.tokenNl();
-        if (nlToken === -1)
-            return null;
-        return nlToken;
-    }
-    /**
-     * @returns {string | null} The BOS (Beginning Of Sequence) token as a string.
-     */
-    getBosString() {
-        const bosToken = this.getBosToken();
-        if (bosToken == null)
-            return null;
-        return this._ctx.getTokenString(bosToken);
-    }
-    /**
-     * @returns {string | null} The EOS (End Of Sequence) token as a string.
-     */
-    getEosString() {
-        const eosToken = this.getEosToken();
-        if (eosToken == null)
-            return null;
-        return this._ctx.getTokenString(eosToken);
-    }
-    /**
-     * @returns {string | null} The NL (New Line) token as a string.
-     */
-    getNlString() {
-        const nlToken = this.getNlToken();
-        if (nlToken == null)
-            return null;
-        return this._ctx.getTokenString(nlToken);
-    }
-    getContextSize() {
-        return this._ctx.getContextSize();
-    }
-    printTimings() {
-        this._ctx.printTimings();
-    }
-    /**
-     * @param {Uint32Array} tokens
-     * @param {object} options
-     * @returns {AsyncGenerator<Token, void>}
-     */
-    async *evaluate(tokens, { temperature = this._model._evaluationOptions.temperature, topK = this._model._evaluationOptions.topK, topP = this._model._evaluationOptions.topP, grammarEvaluationState, repeatPenalty } = {}) {
-        let evalTokens = tokens;
-        if (this._prependTokens.length > 0) {
-            const tokenArray = this._prependTokens.concat(Array.from(tokens));
-            evalTokens = Uint32Array.from(tokenArray);
-            this._prependTokens = [];
-        }
-        if (evalTokens.length === 0)
-            return;
-        // eslint-disable-next-line no-constant-condition
-        while (true) {
-            // Evaluate to get the next token.
-            const nextToken = await this._ctx.eval(evalTokens, removeNullFields({
-                temperature,
-                topK,
-                topP,
-                repeatPenalty: repeatPenalty?.penalty,
-                repeatPenaltyTokens: repeatPenalty?.punishTokens instanceof Function
-                    ? repeatPenalty.punishTokens()
-                    : repeatPenalty?.punishTokens,
-                repeatPenaltyPresencePenalty: repeatPenalty?.presencePenalty,
-                repeatPenaltyFrequencyPenalty: repeatPenalty?.frequencyPenalty,
-                grammarEvaluationState: grammarEvaluationState?._state
-            }));
-            // the assistant finished answering
-            if (nextToken === this._ctx.tokenEos())
-                break;
-            yield nextToken;
-            // Create tokens for the next eval.
-            evalTokens = Uint32Array.from([nextToken]);
-        }
-    }
-}
-//# sourceMappingURL=LlamaContext.js.map

package/dist/llamaEvaluator/LlamaContext.js.map

@@ -1 +0,0 @@
-{"version":3,"file":"LlamaContext.js","sourceRoot":"","sources":["../../src/llamaEvaluator/LlamaContext.ts"],"names":[],"mappings":"AAAA,OAAO,EAAC,gBAAgB,EAAC,MAAM,8BAA8B,CAAC;AAE9D,OAAO,EAAC,YAAY,EAAC,MAAM,gBAAgB,CAAC;AA8D5C,MAAM,OAAO,YAAY;IACJ,MAAM,CAAa;IACnB,IAAI,CAAe;IACnB,WAAW,CAAU;IAC9B,cAAc,CAAU;IAEhC,gBAAgB;IACA,YAAY,CAAgB;IAG5C;;OAEG;IACH,YAAmB,EACf,KAAK,EACL,UAAU,GAAG,IAAI,EACjB,OAAO,EACP,IAAI,GAAG,KAAK,CAAC,eAAe,CAAC,IAAI,EACjC,WAAW,GAAG,KAAK,CAAC,eAAe,CAAC,WAAW,EAC/C,SAAS,GAAG,KAAK,CAAC,eAAe,CAAC,SAAS,EAC3C,SAAS,GAAG,KAAK,CAAC,eAAe,CAAC,SAAS,EAC3C,SAAS,GAAG,KAAK,CAAC,eAAe,CAAC,SAAS,EAC3C,OAAO,GAAG,KAAK,CAAC,eAAe,CAAC,OAAO,EACrB;QAClB,IAAI,CAAC,MAAM,GAAG,KAAK,CAAC;QACpB,IAAI,CAAC,IAAI,GAAG,IAAI,YAAY,CAAC,KAAK,CAAC,MAAM,EAAE,gBAAgB,CAAC;YACxD,IAAI,EAAE,IAAI,IAAI,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC,CAAC,CAAC,SAAS;YACnD,WAAW;YACX,SAAS;YACT,SAAS;YACT,SAAS;YACT,OAAO;SACV,CAAC,CAAC,CAAC;QACJ,IAAI,CAAC,WAAW,GAAG,UAAU,CAAC;QAC9B,IAAI,CAAC,cAAc,GAAG,EAAE,CAAC;QACzB,IAAI,CAAC,YAAY,GAAG,OAAO,CAAC;QAE5B,IAAI,UAAU,EAAE;YACZ,IAAI,CAAC,cAAc,CAAC,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,QAAQ,EAAE,CAAC,CAAC;SACrD;IACL,CAAC;IAEM,MAAM,CAAC,IAAY;QACtB,IAAI,IAAI,KAAK,EAAE;YACX,OAAO,IAAI,WAAW,EAAE,CAAC;QAE7B,OAAO,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC;IAClC,CAAC;IAEM,MAAM,CAAC,MAA6B;QACvC,IAAI,MAAM,CAAC,MAAM,KAAK,CAAC;YACnB,OAAO,EAAE,CAAC;QAEd,IAAI,MAAM,YAAY,WAAW;YAC7B,OAAO,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC;QAEpC,OAAO,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,WAAW,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC;IACtD,CAAC;IAED,IAAW,UAAU;QACjB,OAAO,IAAI,CAAC,WAAW,CAAC;IAC5B,CAAC;IAED;;OAEG;IACI,WAAW;QACd,MAAM,QAAQ,GAAG,IAAI,CAAC,IAAI,CAAC,QAAQ,EAAE,CAAC;QAEtC,IAAI,QAAQ,KAAK,CAAC,CAAC;YACf,OAAO,IAAI,CAAC;QAEhB,OAAO,QAAQ,CAAC;IACpB,CAAC;IAED;;OAEG;IACI,WAAW;QACd,MAAM,QAAQ,GAAG,IAAI,CAAC,IAAI,CAAC,QAAQ,EAAE,CAAC;QAEtC,IAAI,QAAQ,KAAK,CAAC,CAAC;YACf,OAAO,IAAI,CAAC;QAEhB,OAAO,QAAQ,CAAC;IACpB,CAAC;IAED;;OAEG;IACI,UAAU;QACb,MAAM,OAAO,GAAG,IAAI,CAAC,IAAI,CAAC,OAAO,EAAE,CAAC;QAEpC,IAAI,OAAO,KAAK,CAAC,CAAC;YACd,OAAO,IAAI,CAAC;QAEhB,OAAO,OAAO,CAAC;IACnB,CAAC;IAED;;OAEG;IACI,YAAY;QACf,MAAM,QAAQ,GAAG,IAAI,CAAC,WAAW,EAAE,CAAC;QAEpC,IAAI,QAAQ,IAAI,IAAI;YAChB,OAAO,IAAI,CAAC;QAEhB,OAAO,IAAI,CAAC,IAAI,CAAC,cAAc,CAAC,QAAQ,CAAC,CAAC;IAC9C,CAAC;IAED;;OAEG;IACI,YAAY;QACf,MAAM,QAAQ,GAAG,IAAI,CAAC,WAAW,EAAE,CAAC;QAEpC,IAAI,QAAQ,IAAI,IAAI;YAChB,OAAO,IAAI,CAAC;QAEhB,OAAO,IAAI,CAAC,IAAI,CAAC,cAAc,CAAC,QAAQ,CAAC,CAAC;IAC9C,CAAC;IAED;;OAEG;IACI,WAAW;QACd,MAAM,OAAO,GAAG,IAAI,CAAC,UAAU,EAAE,CAAC;QAElC,IAAI,OAAO,IAAI,IAAI;YACf,OAAO,IAAI,CAAC;QAEhB,OAAO,IAAI,CAAC,IAAI,CAAC,cAAc,CAAC,OAAO,CAAC,CAAC;IAC7C,CAAC;IAEM,cAAc;QACjB,OAAO,IAAI,CAAC,IAAI,CAAC,cAAc,EAAE,CAAC;IACtC,CAAC;IAEM,YAAY;QACf,IAAI,CAAC,IAAI,CAAC,YAAY,EAAE,CAAC;IAC7B,CAAC;IAED;;;;OAIG;IACI,KAAK,CAAC,CAAC,QAAQ,CAAC,MAAmB,EAAE,EACxC,WAAW,GAAG,IAAI,CAAC,MAAM,CAAC,kBAAkB,CAAC,WAAW,EACxD,IAAI,GAAG,IAAI,CAAC,MAAM,CAAC,kBAAkB,CAAC,IAAI,EAC1C,IAAI,GAAG,IAAI,CAAC,MAAM,CAAC,kBAAkB,CAAC,IAAI,EAC1C,sBAAsB,EACtB,aAAa,KAIb,EAAE;QACF,IAAI,UAAU,GAAG,MAAM,CAAC;QAExB,IAAI,IAAI,CAAC,cAAc,CAAC,MAAM,GAAG,CAAC,EAAE;YAChC,MAAM,UAAU,GAAY,IAAI,CAAC,cAAc,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC;YAE3E,UAAU,GAAG,WAAW,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC;YAC1C,IAAI,CAAC,cAAc,GAAG,EAAE,CAAC;SAC5B;QAED,IAAI,UAAU,CAAC,MAAM,KAAK,CAAC;YACvB,OAAO;QAEX,iDAAiD;QACjD,OAAO,IAAI,EAAE;YACT,kCAAkC;YAClC,MAAM,SAAS,GAAU,MAAM,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,UAAU,EAAE,gBAAgB,CAAC;gBACvE,WAAW;gBACX,IAAI;gBACJ,IAAI;gBACJ,aAAa,EAAE,aAAa,EAAE,OAAO;gBACrC,mBAAmB,EAAE,aAAa,EAAE,YAAY,YAAY,QAAQ;oBAChE,CAAC,CAAC,aAAa,CAAC,YAAY,EAAE;oBAC9B,CAAC,CAAC,aAAa,EAAE,YAAY;gBACjC,4BAA4B,EAAE,aAAa,EAAE,eAAe;gBAC5D,6BAA6B,EAAE,aAAa,EAAE,gBAAgB;gBAC9D,sBAAsB,EAAE,sBAAsB,EAAE,MAAM;aACzD,CAAC,CAAC,CAAC;YAEJ,mCAAmC;YACnC,IAAI,SAAS,KAAK,IAAI,CAAC,IAAI,CAAC,QAAQ,EAAE;gBAClC,MAAM;YAEV,MAAM,SAAS,CAAC;YAEhB,mCAAmC;YACnC,UAAU,GAAG,WAAW,CAAC,IAAI,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC;SAC9C;IACL,CAAC;CACJ"}

package/dist/llamaEvaluator/LlamaGrammar.d.ts

@@ -1,32 +0,0 @@
-export type LlamaGrammarOptions = {
-    /** GBNF grammar */
-    grammar: string;
-    /** print the grammar to stdout */
-    printGrammar?: boolean;
-    /** Consider any of these texts as EOS for the generated out. Only supported by `LlamaChatSession` */
-    stopStrings?: string[];
-    /** Trim whitespace from the end of the generated text. Only supported by `LlamaChatSession` */
-    trimWhitespaceSuffix?: boolean;
-};
-export declare class LlamaGrammar {
-    private readonly _stopStrings;
-    private readonly _trimWhitespaceSuffix;
-    private readonly _grammarText;
-    /**
-     * > GBNF files are supported.
-     * > More info here: [github:ggerganov/llama.cpp:grammars/README.md](
-     * > https://github.com/ggerganov/llama.cpp/blob/f5fe98d11bdf9e7797bcfb05c0c3601ffc4b9d26/grammars/README.md)
-     * @param {object} options
-     * @param {string} options.grammar - GBNF grammar
-     * @param {string[]} [options.stopStrings] - Consider any of these texts as EOS for the generated out.
-     * Only supported by `LlamaChatSession`
-     * @param {boolean} [options.trimWhitespaceSuffix] - Trim whitespace from the end of the generated text.
-     * Only supported by `LlamaChatSession`
-     * @param {boolean} [options.printGrammar] - print the grammar to stdout
-     */
-    constructor({ grammar, stopStrings, trimWhitespaceSuffix, printGrammar }: LlamaGrammarOptions);
-    get grammar(): string;
-    get stopStrings(): readonly string[];
-    get trimWhitespaceSuffix(): boolean;
-    static getFor(type: "json" | "list" | "arithmetic" | "japanese" | "chess"): Promise<LlamaGrammar>;
-}

package/dist/llamaEvaluator/LlamaGrammar.js.map

@@ -1 +0,0 @@
-{"version":3,"file":"LlamaGrammar.js","sourceRoot":"","sources":["../../src/llamaEvaluator/LlamaGrammar.ts"],"names":[],"mappings":"AAAA,OAAO,IAAI,MAAM,MAAM,CAAC;AACxB,OAAO,EAAE,MAAM,UAAU,CAAC;AAC1B,OAAO,EAAC,iBAAiB,EAAC,MAAM,+BAA+B,CAAC;AAChE,OAAO,EAAC,YAAY,EAAC,MAAM,gBAAgB,CAAC;AAiB5C,MAAM,OAAO,YAAY;IACrB,gBAAgB;IACA,QAAQ,CAAe;IACtB,YAAY,CAAoB;IAChC,qBAAqB,CAAU;IAC/B,YAAY,CAAS;IAEtC;;;;;;;;;;;OAWG;IACH,YAAmB,EACf,OAAO,EAAE,WAAW,GAAG,EAAE,EAAE,oBAAoB,GAAG,KAAK,EAAE,YAAY,GAAG,KAAK,EAC3D;QAClB,IAAI,CAAC,QAAQ,GAAG,IAAI,YAAY,CAAC,OAAO,EAAE;YACtC,YAAY;SACf,CAAC,CAAC;QACH,IAAI,CAAC,YAAY,GAAG,WAAW,IAAI,EAAE,CAAC;QACtC,IAAI,CAAC,qBAAqB,GAAG,oBAAoB,CAAC;QAClD,IAAI,CAAC,YAAY,GAAG,OAAO,CAAC;IAChC,CAAC;IAED,IAAW,OAAO;QACd,OAAO,IAAI,CAAC,YAAY,CAAC;IAC7B,CAAC;IAED,IAAW,WAAW;QAClB,OAAO,IAAI,CAAC,YAAY,CAAC;IAC7B,CAAC;IAED,IAAW,oBAAoB;QAC3B,OAAO,IAAI,CAAC,qBAAqB,CAAC;IACtC,CAAC;IAEM,MAAM,CAAC,KAAK,CAAC,MAAM,CAAC,IAA2D;QAClF,MAAM,cAAc,GAAG,MAAM,iBAAiB,EAAE,CAAC;QAEjD,MAAM,WAAW,GAAG,IAAI,CAAC,IAAI,CAAC,cAAc,EAAE,IAAI,GAAG,OAAO,CAAC,CAAC;QAE9D,IAAI,MAAM,EAAE,CAAC,UAAU,CAAC,WAAW,CAAC,EAAE;YAClC,MAAM,OAAO,GAAG,MAAM,EAAE,CAAC,QAAQ,CAAC,WAAW,EAAE,MAAM,CAAC,CAAC;YACvD,OAAO,IAAI,YAAY,CAAC;gBACpB,OAAO;gBACP,WAAW,EAAE,CAAC,IAAI,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC;gBAC9B,oBAAoB,EAAE,IAAI;aAC7B,CAAC,CAAC;SACN;QAED,MAAM,IAAI,KAAK,CAAC,0BAA0B,IAAI,uBAAuB,cAAc,GAAG,CAAC,CAAC;IAC5F,CAAC;CACJ"}

package/dist/llamaEvaluator/LlamaGrammarEvaluationState.d.ts

@@ -1,14 +0,0 @@
-import { LlamaGrammar } from "./LlamaGrammar.js";
-export type LlamaGrammarEvaluationStateOptions = {
-    grammar: LlamaGrammar;
-};
-export declare class LlamaGrammarEvaluationState {
-    /**
-     * Grammar evaluation state is used to track the model response to determine the next allowed characters for the model to generate.
-     * Create a new grammar evaluation state for every response you generate with the model.
-     * This is only needed when using the `LlamaContext` class directly, as `LlamaChatSession` already handles this for you.
-     * @param {object} options
-     * @param {LlamaGrammar} options.grammar
-     */
-    constructor({ grammar }: LlamaGrammarEvaluationStateOptions);
-}

package/dist/llamaEvaluator/LlamaGrammarEvaluationState.js

@@ -1,16 +0,0 @@
-import { LLAMAGrammarEvaluationState } from "./LlamaBins.js";
-export class LlamaGrammarEvaluationState {
-    /** @internal */
-    _state;
-    /**
-     * Grammar evaluation state is used to track the model response to determine the next allowed characters for the model to generate.
-     * Create a new grammar evaluation state for every response you generate with the model.
-     * This is only needed when using the `LlamaContext` class directly, as `LlamaChatSession` already handles this for you.
-     * @param {object} options
-     * @param {LlamaGrammar} options.grammar
-     */
-    constructor({ grammar }) {
-        this._state = new LLAMAGrammarEvaluationState(grammar._grammar);
-    }
-}
-//# sourceMappingURL=LlamaGrammarEvaluationState.js.map

package/dist/llamaEvaluator/LlamaGrammarEvaluationState.js.map

@@ -1 +0,0 @@
-{"version":3,"file":"LlamaGrammarEvaluationState.js","sourceRoot":"","sources":["../../src/llamaEvaluator/LlamaGrammarEvaluationState.ts"],"names":[],"mappings":"AAAA,OAAO,EAAC,2BAA2B,EAAC,MAAM,gBAAgB,CAAC;AAQ3D,MAAM,OAAO,2BAA2B;IACpC,gBAAgB;IACA,MAAM,CAA8B;IAEpD;;;;;;OAMG;IACH,YAAmB,EAAC,OAAO,EAAqC;QAC5D,IAAI,CAAC,MAAM,GAAG,IAAI,2BAA2B,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC;IACpE,CAAC;CACJ"}

package/dist/llamaEvaluator/LlamaJsonSchemaGrammar.js.map

@@ -1 +0,0 @@
-{"version":3,"file":"LlamaJsonSchemaGrammar.js","sourceRoot":"","sources":["../../src/llamaEvaluator/LlamaJsonSchemaGrammar.ts"],"names":[],"mappings":"AACA,OAAO,EAAC,+BAA+B,EAAC,MAAM,6CAA6C,CAAC;AAC5F,OAAO,EAAC,+BAA+B,EAAC,MAAM,4DAA4D,CAAC;AAC3G,OAAO,EAAC,YAAY,EAAC,MAAM,mBAAmB,CAAC;AAE/C,MAAM,OAAO,sBAAiE,SAAQ,YAAY;IAC7E,OAAO,CAAI;IAE5B,YAAmB,MAAS;QACxB,MAAM,OAAO,GAAG,+BAA+B,CAAC,MAAM,CAAC,CAAC;QAExD,KAAK,CAAC;YACF,OAAO;YACP,WAAW,EAAE,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC;YAC7B,oBAAoB,EAAE,IAAI;SAC7B,CAAC,CAAC;QAEH,IAAI,CAAC,OAAO,GAAG,MAAM,CAAC;IAC1B,CAAC;IAEM,KAAK,CAAC,IAAY;QACrB,MAAM,UAAU,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;QAEpC,+BAA+B,CAAC,UAAU,EAAE,IAAI,CAAC,OAAO,CAAC,CAAC;QAE1D,OAAO,UAAU,CAAC;IACtB,CAAC;CACJ"}

package/dist/llamaEvaluator/LlamaModel.d.ts

@@ -1,123 +0,0 @@
-export type LlamaModelOptions = {
-    /** path to the model on the filesystem */
-    modelPath: string;
-    /**
-     * If null, a random seed will be used
-     * @deprecated use the `seed` option on `LlamaContext` instead
-     * @hidden
-     * */
-    seed?: number | null;
-    /**
-     * text context size
-     * @deprecated use the `contextSize` option on `LlamaContext` instead
-     * @hidden
-     * */
-    contextSize?: number;
-    /**
-     * prompt processing batch size
-     * @deprecated use the `batchSize` option on `LlamaContext` instead
-     * @hidden
-     * */
-    batchSize?: number;
-    /** number of layers to store in VRAM */
-    gpuLayers?: number;
-    /**
-     * number of threads to use to evaluate tokens
-     * @deprecated use the `threads` option on `LlamaContext` instead
-     * @hidden
-     * */
-    threads?: number;
-    /**
-     * Temperature is a hyperparameter that controls the randomness of the generated text.
-     * It affects the probability distribution of the model's output tokens.
-     * A higher temperature (e.g., 1.5) makes the output more random and creative,
-     * while a lower temperature (e.g., 0.5) makes the output more focused, deterministic, and conservative.
-     * The suggested temperature is 0.8, which provides a balance between randomness and determinism.
-     * At the extreme, a temperature of 0 will always pick the most likely next token, leading to identical outputs in each run.
-     *
-     * Set to `0` to disable.
-     * @deprecated use the `temperature` option on `LlamaChatSession`'s `prompt` function or `LlamaContext`'s `evaluate` function instead
-     * @hidden
-     */
-    temperature?: number;
-    /**
-     * Limits the model to consider only the K most likely next tokens for sampling at each step of sequence generation.
-     * An integer number between `1` and the size of the vocabulary.
-     * Set to `0` to disable (which uses the full vocabulary).
-     *
-     * Only relevant when `temperature` is set to a value greater than 0.
-     * @deprecated use the `topK` option on `LlamaChatSession`'s `prompt` function or `LlamaContext`'s `evaluate` function instead
-     * @hidden
-     * */
-    topK?: number;
-    /**
-     * Dynamically selects the smallest set of tokens whose cumulative probability exceeds the threshold P,
-     * and samples the next token only from this set.
-     * A float number between `0` and `1`.
-     * Set to `1` to disable.
-     *
-     * Only relevant when `temperature` is set to a value greater than `0`.
-     * @deprecated use the `topP` option on `LlamaChatSession`'s `prompt` function or `LlamaContext`'s `evaluate` function instead
-     * @hidden
-     */
-    topP?: number;
-    /**
-     * the llama_eval() call computes all logits, not just the last one
-     * @deprecated use the `logitsAll` option on `LlamaContext` instead
-     * @hidden
-     */
-    logitsAll?: boolean;
-    /** only load the vocabulary, no weights */
-    vocabOnly?: boolean;
-    /** use mmap if possible */
-    useMmap?: boolean;
-    /** force system to keep model in RAM */
-    useMlock?: boolean;
-    /**
-     * embedding mode only
-     * @deprecated use the `embedding` option on `LlamaContext` instead
-     * @hidden
-     */
-    embedding?: boolean;
-};
-export declare class LlamaModel {
-    /**
-     * > options source:
-     * > [github:ggerganov/llama.cpp/llama.h](
-     * > https://github.com/ggerganov/llama.cpp/blob/b5ffb2849d23afe73647f68eec7b68187af09be6/llama.h#L102) (`struct llama_context_params`)
-     * @param {object} options
-     * @param {string} options.modelPath - path to the model on the filesystem
-     * @param {number | null} [options.seed] - If null, a random seed will be used
-     * @param {number} [options.contextSize] - text context size
-     * @param {number} [options.batchSize] - prompt processing batch size
-     * @param {number} [options.gpuLayers] - number of layers to store in VRAM
-     * @param {number} [options.threads] - number of threads to use to evaluate tokens
-     * @param {number} [options.temperature] - Temperature is a hyperparameter that controls the randomness of the generated text.
-     * It affects the probability distribution of the model's output tokens.
-     * A higher temperature (e.g., 1.5) makes the output more random and creative,
-     * while a lower temperature (e.g., 0.5) makes the output more focused, deterministic, and conservative.
-     * The suggested temperature is 0.8, which provides a balance between randomness and determinism.
-     * At the extreme, a temperature of 0 will always pick the most likely next token, leading to identical outputs in each run.
-     *
-     * Set to `0` to disable.
-     * @param {number} [options.topK] - Limits the model to consider only the K most likely next tokens for sampling at each step of
-     * sequence generation.
-     * An integer number between `1` and the size of the vocabulary.
-     * Set to `0` to disable (which uses the full vocabulary).
-     *
-     * Only relevant when `temperature` is set to a value greater than 0.
-     * @param {number} [options.topP] - Dynamically selects the smallest set of tokens whose cumulative probability exceeds the threshold P,
-     * and samples the next token only from this set.
-     * A float number between `0` and `1`.
-     * Set to `1` to disable.
-     *
-     * Only relevant when `temperature` is set to a value greater than `0`.
-     * @param {boolean} [options.logitsAll] - the llama_eval() call computes all logits, not just the last one
-     * @param {boolean} [options.vocabOnly] - only load the vocabulary, no weights
-     * @param {boolean} [options.useMmap] - use mmap if possible
-     * @param {boolean} [options.useMlock] - force system to keep model in RAM
-     * @param {boolean} [options.embedding] - embedding mode only
-     */
-    constructor({ modelPath, seed, contextSize, batchSize, gpuLayers, threads, temperature, topK, topP, logitsAll, vocabOnly, useMmap, useMlock, embedding }: LlamaModelOptions);
-    static get systemInfo(): string;
-}

package/dist/llamaEvaluator/LlamaModel.js

@@ -1,74 +0,0 @@
-import process from "process";
-import path from "path";
-import { removeNullFields } from "../utils/removeNullFields.js";
-import { llamaCppNode, LLAMAModel } from "./LlamaBins.js";
-export class LlamaModel {
-    /** @internal */
-    _model;
-    /** @internal */
-    _contextOptions;
-    /** @internal */
-    _evaluationOptions;
-    /**
-     * > options source:
-     * > [github:ggerganov/llama.cpp/llama.h](
-     * > https://github.com/ggerganov/llama.cpp/blob/b5ffb2849d23afe73647f68eec7b68187af09be6/llama.h#L102) (`struct llama_context_params`)
-     * @param {object} options
-     * @param {string} options.modelPath - path to the model on the filesystem
-     * @param {number | null} [options.seed] - If null, a random seed will be used
-     * @param {number} [options.contextSize] - text context size
-     * @param {number} [options.batchSize] - prompt processing batch size
-     * @param {number} [options.gpuLayers] - number of layers to store in VRAM
-     * @param {number} [options.threads] - number of threads to use to evaluate tokens
-     * @param {number} [options.temperature] - Temperature is a hyperparameter that controls the randomness of the generated text.
-     * It affects the probability distribution of the model's output tokens.
-     * A higher temperature (e.g., 1.5) makes the output more random and creative,
-     * while a lower temperature (e.g., 0.5) makes the output more focused, deterministic, and conservative.
-     * The suggested temperature is 0.8, which provides a balance between randomness and determinism.
-     * At the extreme, a temperature of 0 will always pick the most likely next token, leading to identical outputs in each run.
-     *
-     * Set to `0` to disable.
-     * @param {number} [options.topK] - Limits the model to consider only the K most likely next tokens for sampling at each step of
-     * sequence generation.
-     * An integer number between `1` and the size of the vocabulary.
-     * Set to `0` to disable (which uses the full vocabulary).
-     *
-     * Only relevant when `temperature` is set to a value greater than 0.
-     * @param {number} [options.topP] - Dynamically selects the smallest set of tokens whose cumulative probability exceeds the threshold P,
-     * and samples the next token only from this set.
-     * A float number between `0` and `1`.
-     * Set to `1` to disable.
-     *
-     * Only relevant when `temperature` is set to a value greater than `0`.
-     * @param {boolean} [options.logitsAll] - the llama_eval() call computes all logits, not just the last one
-     * @param {boolean} [options.vocabOnly] - only load the vocabulary, no weights
-     * @param {boolean} [options.useMmap] - use mmap if possible
-     * @param {boolean} [options.useMlock] - force system to keep model in RAM
-     * @param {boolean} [options.embedding] - embedding mode only
-     */
-    constructor({ modelPath, seed = null, contextSize = 1024 * 4, batchSize, gpuLayers, threads = 6, temperature = 0, topK = 40, topP = 0.95, logitsAll, vocabOnly, useMmap, useMlock, embedding }) {
-        this._model = new LLAMAModel(path.resolve(process.cwd(), modelPath), removeNullFields({
-            gpuLayers,
-            vocabOnly,
-            useMmap,
-            useMlock
-        }));
-        this._contextOptions = {
-            seed,
-            contextSize,
-            batchSize,
-            logitsAll,
-            embedding,
-            threads
-        };
-        this._evaluationOptions = {
-            temperature,
-            topK,
-            topP
-        };
-    }
-    static get systemInfo() {
-        return llamaCppNode.systemInfo();
-    }
-}
-//# sourceMappingURL=LlamaModel.js.map

package/dist/llamaEvaluator/LlamaModel.js.map

@@ -1 +0,0 @@
-{"version":3,"file":"LlamaModel.js","sourceRoot":"","sources":["../../src/llamaEvaluator/LlamaModel.ts"],"names":[],"mappings":"AAAA,OAAO,OAAO,MAAM,SAAS,CAAC;AAC9B,OAAO,IAAI,MAAM,MAAM,CAAC;AACxB,OAAO,EAAC,gBAAgB,EAAC,MAAM,8BAA8B,CAAC;AAC9D,OAAO,EAAC,YAAY,EAAE,UAAU,EAAC,MAAM,gBAAgB,CAAC;AAmGxD,MAAM,OAAO,UAAU;IACnB,gBAAgB;IACA,MAAM,CAAa;IAEnC,gBAAgB;IACA,eAAe,CAO7B;IAEF,gBAAgB;IACA,kBAAkB,CAIhC;IAEF;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;OAoCG;IACH,YAAmB,EACf,SAAS,EAAE,IAAI,GAAG,IAAI,EAAE,WAAW,GAAG,IAAI,GAAG,CAAC,EAAE,SAAS,EAAE,SAAS,EACpE,OAAO,GAAG,CAAC,EAAE,WAAW,GAAG,CAAC,EAAE,IAAI,GAAG,EAAE,EAAE,IAAI,GAAG,IAAI,EAAE,SAAS,EAAE,SAAS,EAAE,OAAO,EAAE,QAAQ,EAAE,SAAS,EACxF;QAChB,IAAI,CAAC,MAAM,GAAG,IAAI,UAAU,CAAC,IAAI,CAAC,OAAO,CAAC,OAAO,CAAC,GAAG,EAAE,EAAE,SAAS,CAAC,EAAE,gBAAgB,CAAC;YAClF,SAAS;YACT,SAAS;YACT,OAAO;YACP,QAAQ;SACX,CAAC,CAAC,CAAC;QAEJ,IAAI,CAAC,eAAe,GAAG;YACnB,IAAI;YACJ,WAAW;YACX,SAAS;YACT,SAAS;YACT,SAAS;YACT,OAAO;SACV,CAAC;QAEF,IAAI,CAAC,kBAAkB,GAAG;YACtB,WAAW;YACX,IAAI;YACJ,IAAI;SACP,CAAC;IACN,CAAC;IAEM,MAAM,KAAK,UAAU;QACxB,OAAO,YAAY,CAAC,UAAU,EAAE,CAAC;IACrC,CAAC;CACJ"}

package/dist/utils/binariesGithubRelease.js.map

@@ -1 +0,0 @@
-{"version":3,"file":"binariesGithubRelease.js","sourceRoot":"","sources":["../../src/utils/binariesGithubRelease.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,MAAM,UAAU,CAAC;AAC1B,OAAO,EAAC,yBAAyB,EAAC,MAAM,cAAc,CAAC;AAMvD,MAAM,CAAC,KAAK,UAAU,wBAAwB;IAC1C,MAAM,qBAAqB,GAA8B,MAAM,EAAE,CAAC,QAAQ,CAAC,yBAAyB,CAAC,CAAC;IAEtG,OAAO,qBAAqB,CAAC,OAAO,CAAC;AACzC,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,wBAAwB,CAAC,OAA6C;IACxF,MAAM,yBAAyB,GAA8B;QACzD,OAAO,EAAE,OAAO;KACnB,CAAC;IAEF,MAAM,EAAE,CAAC,SAAS,CAAC,yBAAyB,EAAE,yBAAyB,EAAE;QACrE,MAAM,EAAE,CAAC;KACZ,CAAC,CAAC;AACP,CAAC"}

package/dist/utils/clearLlamaBuild.d.ts

@@ -1 +0,0 @@
-export declare function clearLlamaBuild(): Promise<void>;

package/dist/utils/clearLlamaBuild.js

@@ -1,12 +0,0 @@
-import path from "path";
-import fs from "fs-extra";
-import { llamaDirectory } from "../config.js";
-import { clearTempFolder } from "./clearTempFolder.js";
-export async function clearLlamaBuild() {
-    await fs.remove(path.join(llamaDirectory, "Debug"));
-    await fs.remove(path.join(llamaDirectory, "Release"));
-    await fs.remove(path.join(llamaDirectory, "compile_commands.json"));
-    await fs.remove(path.join(llamaDirectory, "build"));
-    await clearTempFolder();
-}
-//# sourceMappingURL=clearLlamaBuild.js.map

package/dist/utils/clearLlamaBuild.js.map

@@ -1 +0,0 @@
-{"version":3,"file":"clearLlamaBuild.js","sourceRoot":"","sources":["../../src/utils/clearLlamaBuild.ts"],"names":[],"mappings":"AAAA,OAAO,IAAI,MAAM,MAAM,CAAC;AACxB,OAAO,EAAE,MAAM,UAAU,CAAC;AAC1B,OAAO,EAAC,cAAc,EAAC,MAAM,cAAc,CAAC;AAC5C,OAAO,EAAC,eAAe,EAAC,MAAM,sBAAsB,CAAC;AAErD,MAAM,CAAC,KAAK,UAAU,eAAe;IACjC,MAAM,EAAE,CAAC,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,cAAc,EAAE,OAAO,CAAC,CAAC,CAAC;IACpD,MAAM,EAAE,CAAC,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,cAAc,EAAE,SAAS,CAAC,CAAC,CAAC;IACtD,MAAM,EAAE,CAAC,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,cAAc,EAAE,uBAAuB,CAAC,CAAC,CAAC;IACpE,MAAM,EAAE,CAAC,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,cAAc,EAAE,OAAO,CAAC,CAAC,CAAC;IAEpD,MAAM,eAAe,EAAE,CAAC;AAC5B,CAAC"}

package/dist/utils/cloneLlamaCppRepo.d.ts

@@ -1,2 +0,0 @@
-export declare function cloneLlamaCppRepo(githubOwner: string, githubRepo: string, tag: string, useBundles?: boolean): Promise<void>;
-export declare function getClonedLlamaCppRepoReleaseTag(): Promise<string | null>;

package/dist/utils/cloneLlamaCppRepo.js

@@ -1,102 +0,0 @@
-import simpleGit from "simple-git";
-import cliProgress from "cli-progress";
-import chalk from "chalk";
-import fs from "fs-extra";
-import { llamaCppDirectory, llamaCppDirectoryTagFilePath } from "../config.js";
-import { getGitBundlePathForRelease } from "./gitReleaseBundles.js";
-export async function cloneLlamaCppRepo(githubOwner, githubRepo, tag, useBundles = true) {
-    const gitBundleForTag = !useBundles ? null : await getGitBundlePathForRelease(githubOwner, githubRepo, tag);
-    const remoteGitUrl = `https://github.com/${githubOwner}/${githubRepo}.git`;
-    async function withGitCloneProgress(cloneName, callback) {
-        const progressBar = new cliProgress.Bar({
-            clearOnComplete: false,
-            hideCursor: true,
-            autopadding: true,
-            format: `${chalk.bold("Clone {repo}")}  ${chalk.yellow("{percentage}%")} ${chalk.cyan("{bar}")} ${chalk.grey("{eta_formatted}")}`
-        }, cliProgress.Presets.shades_classic);
-        progressBar.start(100, 0, {
-            speed: "",
-            repo: `${githubOwner}/${githubRepo} (${cloneName})`
-        });
-        const gitWithCloneProgress = simpleGit({
-            progress({ progress, total, processed }) {
-                const totalProgress = (processed / 100) + (progress / total);
-                progressBar.update(Math.floor(totalProgress * 10000) / 100);
-            }
-        });
-        try {
-            const res = await callback(gitWithCloneProgress);
-            progressBar.update(100);
-            return res;
-        }
-        finally {
-            progressBar.stop();
-        }
-    }
-    if (gitBundleForTag != null) {
-        try {
-            await withGitCloneProgress("local bundle", async (gitWithCloneProgress) => {
-                await gitWithCloneProgress.clone(gitBundleForTag, llamaCppDirectory, {
-                    "--quiet": null
-                });
-                await simpleGit(llamaCppDirectory).removeRemote("origin");
-            });
-            return;
-        }
-        catch (err) {
-            await fs.remove(llamaCppDirectory);
-            await fs.remove(llamaCppDirectoryTagFilePath);
-            console.error("Failed to clone git bundle, cloning from GitHub instead", err);
-            printCloneErrorHelp(String(err));
-        }
-    }
-    try {
-        await withGitCloneProgress("GitHub", async (gitWithCloneProgress) => {
-            await gitWithCloneProgress.clone(remoteGitUrl, llamaCppDirectory, {
-                "--depth": 1,
-                "--branch": tag,
-                "--quiet": null
-            });
-        });
-    }
-    catch (err) {
-        printCloneErrorHelp(String(err));
-        throw err;
-    }
-    try {
-        const clonedLlamaCppRepoTagJson = {
-            tag
-        };
-        await fs.writeJson(llamaCppDirectoryTagFilePath, clonedLlamaCppRepoTagJson, {
-            spaces: 4
-        });
-    }
-    catch (err) {
-        console.error("Failed to write llama.cpp tag file", err);
-        throw err;
-    }
-}
-function printCloneErrorHelp(error) {
-    // This error happens with some docker images where the current user is different
-    // from the owner of the files due to mounting a volume.
-    // In such cases, print a helpful message to help the user resolve the issue.
-    if (error.toLowerCase().includes("detected dubious ownership in repository"))
-        console.info("\n" +
-            chalk.grey("[node-llama-cpp]") + chalk.yellow(" To fix this issue, try running this command to fix it for the current module directory:") + "\n" +
-            'git config --global --add safe.directory "' + llamaCppDirectory + '"\n\n' +
-            chalk.yellow("Or run this command to fix it everywhere:") + "\n" +
-            'git config --global --add safe.directory "*"');
-}
-export async function getClonedLlamaCppRepoReleaseTag() {
-    if (!(await fs.pathExists(llamaCppDirectoryTagFilePath)))
-        return null;
-    try {
-        const clonedLlamaCppRepoTagJson = await fs.readJson(llamaCppDirectoryTagFilePath);
-        return clonedLlamaCppRepoTagJson.tag;
-    }
-    catch (err) {
-        console.error("Failed to read llama.cpp tag file", err);
-        return null;
-    }
-}
-//# sourceMappingURL=cloneLlamaCppRepo.js.map

package/dist/utils/cloneLlamaCppRepo.js.map

@@ -1 +0,0 @@
-{"version":3,"file":"cloneLlamaCppRepo.js","sourceRoot":"","sources":["../../src/utils/cloneLlamaCppRepo.ts"],"names":[],"mappings":"AAAA,OAAO,SAAsB,MAAM,YAAY,CAAC;AAChD,OAAO,WAAW,MAAM,cAAc,CAAC;AACvC,OAAO,KAAK,MAAM,OAAO,CAAC;AAC1B,OAAO,EAAE,MAAM,UAAU,CAAC;AAC1B,OAAO,EAAC,iBAAiB,EAAE,4BAA4B,EAAC,MAAM,cAAc,CAAC;AAC7E,OAAO,EAAC,0BAA0B,EAAC,MAAM,wBAAwB,CAAC;AAOlE,MAAM,CAAC,KAAK,UAAU,iBAAiB,CAAC,WAAmB,EAAE,UAAkB,EAAE,GAAW,EAAE,aAAsB,IAAI;IACpH,MAAM,eAAe,GAAG,CAAC,UAAU,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,MAAM,0BAA0B,CAAC,WAAW,EAAE,UAAU,EAAE,GAAG,CAAC,CAAC;IAC5G,MAAM,YAAY,GAAG,sBAAsB,WAAW,IAAI,UAAU,MAAM,CAAC;IAE3E,KAAK,UAAU,oBAAoB,CAAI,SAAiB,EAAE,QAAyD;QAC/G,MAAM,WAAW,GAAG,IAAI,WAAW,CAAC,GAAG,CAAC;YACpC,eAAe,EAAE,KAAK;YACtB,UAAU,EAAE,IAAI;YAChB,WAAW,EAAE,IAAI;YACjB,MAAM,EAAE,GAAG,KAAK,CAAC,IAAI,CAAC,cAAc,CAAC,KAAK,KAAK,CAAC,MAAM,CAAC,eAAe,CAAC,IAAI,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,IAAI,KAAK,CAAC,IAAI,CAAC,iBAAiB,CAAC,EAAE;SACpI,EAAE,WAAW,CAAC,OAAO,CAAC,cAAc,CAAC,CAAC;QAEvC,WAAW,CAAC,KAAK,CAAC,GAAG,EAAE,CAAC,EAAE;YACtB,KAAK,EAAE,EAAE;YACT,IAAI,EAAE,GAAG,WAAW,IAAI,UAAU,KAAK,SAAS,GAAG;SACtD,CAAC,CAAC;QAEH,MAAM,oBAAoB,GAAG,SAAS,CAAC;YACnC,QAAQ,CAAC,EAAC,QAAQ,EAAE,KAAK,EAAE,SAAS,EAAC;gBACjC,MAAM,aAAa,GAAG,CAAC,SAAS,GAAG,GAAG,CAAC,GAAG,CAAC,QAAQ,GAAG,KAAK,CAAC,CAAC;gBAE7D,WAAW,CAAC,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,aAAa,GAAG,KAAK,CAAC,GAAG,GAAG,CAAC,CAAC;YAChE,CAAC;SACJ,CAAC,CAAC;QAEH,IAAI;YACA,MAAM,GAAG,GAAG,MAAM,QAAQ,CAAC,oBAAoB,CAAC,CAAC;YAEjD,WAAW,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;YAExB,OAAO,GAAG,CAAC;SACd;gBAAS;YACN,WAAW,CAAC,IAAI,EAAE,CAAC;SACtB;IACL,CAAC;IAED,IAAI,eAAe,IAAI,IAAI,EAAE;QACzB,IAAI;YACA,MAAM,oBAAoB,CAAC,cAAc,EAAE,KAAK,EAAE,oBAAoB,EAAE,EAAE;gBACtE,MAAM,oBAAoB,CAAC,KAAK,CAAC,eAAe,EAAE,iBAAiB,EAAE;oBACjE,SAAS,EAAE,IAAI;iBAClB,CAAC,CAAC;gBAEH,MAAM,SAAS,CAAC,iBAAiB,CAAC,CAAC,YAAY,CAAC,QAAQ,CAAC,CAAC;YAC9D,CAAC,CAAC,CAAC;YACH,OAAO;SACV;QAAC,OAAO,GAAG,EAAE;YACV,MAAM,EAAE,CAAC,MAAM,CAAC,iBAAiB,CAAC,CAAC;YACnC,MAAM,EAAE,CAAC,MAAM,CAAC,4BAA4B,CAAC,CAAC;YAC9C,OAAO,CAAC,KAAK,CAAC,yDAAyD,EAAE,GAAG,CAAC,CAAC;YAE9E,mBAAmB,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC;SACpC;KACJ;IAED,IAAI;QACA,MAAM,oBAAoB,CAAC,QAAQ,EAAE,KAAK,EAAE,oBAAoB,EAAE,EAAE;YAChE,MAAM,oBAAoB,CAAC,KAAK,CAAC,YAAY,EAAE,iBAAiB,EAAE;gBAC9D,SAAS,EAAE,CAAC;gBACZ,UAAU,EAAE,GAAG;gBACf,SAAS,EAAE,IAAI;aAClB,CAAC,CAAC;QACP,CAAC,CAAC,CAAC;KACN;IAAC,OAAO,GAAG,EAAE;QACV,mBAAmB,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC;QAEjC,MAAM,GAAG,CAAC;KACb;IAED,IAAI;QACA,MAAM,yBAAyB,GAA8B;YACzD,GAAG;SACN,CAAC;QAEF,MAAM,EAAE,CAAC,SAAS,CAAC,4BAA4B,EAAE,yBAAyB,EAAE;YACxE,MAAM,EAAE,CAAC;SACZ,CAAC,CAAC;KACN;IAAC,OAAO,GAAG,EAAE;QACV,OAAO,CAAC,KAAK,CAAC,oCAAoC,EAAE,GAAG,CAAC,CAAC;QAEzD,MAAM,GAAG,CAAC;KACb;AACL,CAAC;AAED,SAAS,mBAAmB,CAAC,KAAa;IACtC,iFAAiF;IACjF,wDAAwD;IACxD,6EAA6E;IAC7E,IAAI,KAAK,CAAC,WAAW,EAAE,CAAC,QAAQ,CAAC,0CAA0C,CAAC;QACxE,OAAO,CAAC,IAAI,CAAC,IAAI;YACb,KAAK,CAAC,IAAI,CAAC,kBAAkB,CAAC,GAAG,KAAK,CAAC,MAAM,CAAC,0FAA0F,CAAC,GAAG,IAAI;YAChJ,4CAA4C,GAAG,iBAAiB,GAAG,OAAO;YAC1E,KAAK,CAAC,MAAM,CAAC,2CAA2C,CAAC,GAAG,IAAI;YAChE,8CAA8C,CACjD,CAAC;AACV,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,+BAA+B;IACjD,IAAI,CAAC,CAAC,MAAM,EAAE,CAAC,UAAU,CAAC,4BAA4B,CAAC,CAAC;QACpD,OAAO,IAAI,CAAC;IAEhB,IAAI;QACA,MAAM,yBAAyB,GAA8B,MAAM,EAAE,CAAC,QAAQ,CAAC,4BAA4B,CAAC,CAAC;QAE7G,OAAO,yBAAyB,CAAC,GAAG,CAAC;KACxC;IAAC,OAAO,GAAG,EAAE;QACV,OAAO,CAAC,KAAK,CAAC,mCAAmC,EAAE,GAAG,CAAC,CAAC;QACxD,OAAO,IAAI,CAAC;KACf;AACL,CAAC"}

package/dist/utils/compileLLamaCpp.d.ts

@@ -1,8 +0,0 @@
-export declare function compileLlamaCpp({ arch, nodeTarget, setUsedBinFlag: setUsedBinFlagArg, metal, cuda }: {
-    arch?: string;
-    nodeTarget?: string;
-    setUsedBinFlag?: boolean;
-    metal?: boolean;
-    cuda?: boolean;
-}): Promise<void>;
-export declare function getCompiledLlamaCppBinaryPath(): Promise<string | null>;