modelfusion 0.106.0 → 0.108.0

package/model-provider/llamacpp/{LlamaCppTextGenerationModel.d.ts → LlamaCppCompletionModel.d.ts}

@@ -8,7 +8,7 @@ import { PromptTemplateTextStreamingModel } from "../../model-function/generate-
 import { TextGenerationModelSettings, TextStreamingModel } from "../../model-function/generate-text/TextGenerationModel.js";
 import { TextGenerationPromptTemplate } from "../../model-function/generate-text/TextGenerationPromptTemplate.js";
 import { LlamaCppTokenizer } from "./LlamaCppTokenizer.js";
-export interface LlamaCppTextGenerationModelSettings<CONTEXT_WINDOW_SIZE extends number | undefined> extends TextGenerationModelSettings {
+export interface LlamaCppCompletionModelSettings<CONTEXT_WINDOW_SIZE extends number | undefined> extends TextGenerationModelSettings {
     api?: ApiConfiguration;
     /**
      * Specify the context window size of the model that you have loaded in your
@@ -16,26 +16,113 @@ export interface LlamaCppTextGenerationModelSettings<CONTEXT_WINDOW_SIZE extends
      */
     contextWindowSize?: CONTEXT_WINDOW_SIZE;
     /**
-     * Save the prompt and generation for avoid reprocess entire prompt if a part of this isn't change (default: false)
+     * Adjust the randomness of the generated text (default: 0.8).
      */
-    cachePrompt?: boolean;
     temperature?: number;
+    /**
+     * Limit the next token selection to the K most probable tokens (default: 40).
+     */
     topK?: number;
+    /**
+     * Limit the next token selection to a subset of tokens with a cumulative probability above a threshold P (default: 0.95).
+     */
     topP?: number;
+    /**
+     * The minimum probability for a token to be considered, relative to the probability of the most likely token (default: 0.05).
+     */
+    minP?: number;
+    /**
+     * Specify the number of tokens from the prompt to retain when the context size is exceeded
+     * and tokens need to be discarded. By default, this value is set to 0 (meaning no tokens
+     * are kept). Use -1 to retain all tokens from the prompt.
+     */
     nKeep?: number;
+    /**
+     * Enable tail free sampling with parameter z (default: 1.0, 1.0 = disabled).
+     */
     tfsZ?: number;
+    /**
+     * Enable locally typical sampling with parameter p (default: 1.0, 1.0 = disabled).
+     */
     typicalP?: number;
+    /**
+     * Control the repetition of token sequences in the generated text (default: 1.1).
+     */
     repeatPenalty?: number;
+    /**
+     * Last n tokens to consider for penalizing repetition (default: 64, 0 = disabled, -1 = ctx-size).
+     */
     repeatLastN?: number;
+    /**
+     * Penalize newline tokens when applying the repeat penalty (default: true).
+     */
     penalizeNl?: boolean;
+    /**
+     * Repeat alpha presence penalty (default: 0.0, 0.0 = disabled).
+     */
+    presencePenalty?: number;
+    /**
+     * Repeat alpha frequency penalty (default: 0.0, 0.0 = disabled).
+     */
+    frequencyPenalty?: number;
+    /**
+     * This will replace the prompt for the purpose of the penalty evaluation.
+     * Can be either null, a string or an array of numbers representing tokens
+     * (default: null = use the original prompt).
+     */
+    penaltyPrompt?: string | number[];
+    /**
+     * Enable Mirostat sampling, controlling perplexity during text generation
+     * (default: 0, 0 = disabled, 1 = Mirostat, 2 = Mirostat 2.0).
+     */
     mirostat?: number;
+    /**
+     * Set the Mirostat target entropy, parameter tau (default: 5.0).
+     */
     mirostatTau?: number;
+    /**
+     * Set the Mirostat learning rate, parameter eta (default: 0.1).
+     */
     mirostatEta?: number;
+    /**
+     * Set grammar for grammar-based sampling (default: no grammar)
+     *
+     * @see https://github.com/ggerganov/llama.cpp/blob/master/grammars/README.md
+     */
+    grammar?: string;
+    /**
+     * Set the random number generator (RNG) seed
+     * (default: -1, -1 = random seed).
+     */
     seed?: number;
+    /**
+     * Ignore end of stream token and continue generating (default: false).
+     */
     ignoreEos?: boolean;
+    /**
+     * Modify the likelihood of a token appearing in the generated text completion.
+     * For example, use "logit_bias": [[15043,1.0]] to increase the likelihood of the token
+     * 'Hello', or "logit_bias": [[15043,-1.0]] to decrease its likelihood.
+     * Setting the value to false, "logit_bias": [[15043,false]] ensures that the token Hello is
+     * never produced (default: []).
+     */
     logitBias?: Array<[number, number | false]>;
+    /**
+     * If greater than 0, the response also contains the probabilities of top N tokens
+     * for each generated token (default: 0)
+     */
+    nProbs?: number;
+    /**
+     * Save the prompt and generation for avoid reprocess entire prompt if a part of this isn't change (default: false)
+     */
+    cachePrompt?: boolean;
+    /**
+     * Assign the completion task to an specific slot.
+     * If is -1 the task will be assigned to a Idle slot (default: -1)
+     */
+    slotId?: number;
 }
-export interface LlamaCppTextGenerationPrompt {
+export interface LlamaCppCompletionPrompt {
     /**
      * Text prompt. Images can be included through references such as `[img-ID]`, e.g. `[img-1]`.
      */
@@ -45,18 +132,18 @@ export interface LlamaCppTextGenerationPrompt {
      */
     images?: Record<number, string>;
 }
-export declare class LlamaCppTextGenerationModel<CONTEXT_WINDOW_SIZE extends number | undefined> extends AbstractModel<LlamaCppTextGenerationModelSettings<CONTEXT_WINDOW_SIZE>> implements TextStreamingModel<LlamaCppTextGenerationPrompt, LlamaCppTextGenerationModelSettings<CONTEXT_WINDOW_SIZE>> {
-    constructor(settings?: LlamaCppTextGenerationModelSettings<CONTEXT_WINDOW_SIZE>);
+export declare class LlamaCppCompletionModel<CONTEXT_WINDOW_SIZE extends number | undefined> extends AbstractModel<LlamaCppCompletionModelSettings<CONTEXT_WINDOW_SIZE>> implements TextStreamingModel<LlamaCppCompletionPrompt, LlamaCppCompletionModelSettings<CONTEXT_WINDOW_SIZE>> {
+    constructor(settings?: LlamaCppCompletionModelSettings<CONTEXT_WINDOW_SIZE>);
     readonly provider = "llamacpp";
     get modelName(): null;
     get contextWindowSize(): CONTEXT_WINDOW_SIZE;
     readonly tokenizer: LlamaCppTokenizer;
-    callAPI<RESPONSE>(prompt: LlamaCppTextGenerationPrompt, options: {
-        responseFormat: LlamaCppTextGenerationResponseFormatType<RESPONSE>;
+    callAPI<RESPONSE>(prompt: LlamaCppCompletionPrompt, options: {
+        responseFormat: LlamaCppCompletionResponseFormatType<RESPONSE>;
     } & FunctionOptions): Promise<RESPONSE>;
-    get settingsForEvent(): Partial<LlamaCppTextGenerationModelSettings<CONTEXT_WINDOW_SIZE>>;
-    countPromptTokens(prompt: LlamaCppTextGenerationPrompt): Promise<number>;
-    doGenerateTexts(prompt: LlamaCppTextGenerationPrompt, options?: FunctionOptions): Promise<{
+    get settingsForEvent(): Partial<LlamaCppCompletionModelSettings<CONTEXT_WINDOW_SIZE>>;
+    countPromptTokens(prompt: LlamaCppCompletionPrompt): Promise<number>;
+    doGenerateTexts(prompt: LlamaCppCompletionPrompt, options?: FunctionOptions): Promise<{
         response: {
             model: string;
             stop: true;
@@ -81,11 +168,11 @@ export declare class LlamaCppTextGenerationModel<CONTEXT_WINDOW_SIZE extends num
                 presence_penalty: number;
                 repeat_last_n: number;
                 repeat_penalty: number;
-                temp: number;
                 tfs_z: number;
                 top_k: number;
                 top_p: number;
                 typical_p: number;
+                temperature?: number | undefined;
             };
             stopped_eos: boolean;
             stopped_limit: boolean;
@@ -96,10 +183,10 @@ export declare class LlamaCppTextGenerationModel<CONTEXT_WINDOW_SIZE extends num
                 predicted_n: number;
                 predicted_per_second: number | null;
                 predicted_per_token_ms: number | null;
-                prompt_ms: number | null;
                 prompt_n: number;
                 prompt_per_second: number | null;
                 prompt_per_token_ms: number | null;
+                prompt_ms?: number | null | undefined;
             };
             tokens_cached: number;
             tokens_evaluated: number;
@@ -116,7 +203,7 @@ export declare class LlamaCppTextGenerationModel<CONTEXT_WINDOW_SIZE extends num
             totalTokens: number;
         };
     }>;
-    doStreamText(prompt: LlamaCppTextGenerationPrompt, options?: FunctionOptions): Promise<AsyncIterable<Delta<{
+    doStreamText(prompt: LlamaCppCompletionPrompt, options?: FunctionOptions): Promise<AsyncIterable<Delta<{
         model: string;
         stop: true;
         content: string;
@@ -140,11 +227,11 @@ export declare class LlamaCppTextGenerationModel<CONTEXT_WINDOW_SIZE extends num
             presence_penalty: number;
             repeat_last_n: number;
             repeat_penalty: number;
-            temp: number;
             tfs_z: number;
             top_k: number;
             top_p: number;
             typical_p: number;
+            temperature?: number | undefined;
         };
         stopped_eos: boolean;
         stopped_limit: boolean;
@@ -155,10 +242,10 @@ export declare class LlamaCppTextGenerationModel<CONTEXT_WINDOW_SIZE extends num
             predicted_n: number;
             predicted_per_second: number | null;
             predicted_per_token_ms: number | null;
-            prompt_ms: number | null;
             prompt_n: number;
             prompt_per_second: number | null;
             prompt_per_token_ms: number | null;
+            prompt_ms?: number | null | undefined;
         };
         tokens_cached: number;
         tokens_evaluated: number;
@@ -169,16 +256,16 @@ export declare class LlamaCppTextGenerationModel<CONTEXT_WINDOW_SIZE extends num
         content: string;
     }>>>;
     extractTextDelta(delta: unknown): string;
-    withTextPrompt(): PromptTemplateTextStreamingModel<string, LlamaCppTextGenerationPrompt, LlamaCppTextGenerationModelSettings<CONTEXT_WINDOW_SIZE>, this>;
+    withTextPrompt(): PromptTemplateTextStreamingModel<string, LlamaCppCompletionPrompt, LlamaCppCompletionModelSettings<CONTEXT_WINDOW_SIZE>, this>;
     /**
      * Maps the prompt for a text version of the Llama.cpp prompt template (without image support).
      */
-    withTextPromptTemplate<INPUT_PROMPT>(promptTemplate: TextGenerationPromptTemplate<INPUT_PROMPT, string>): PromptTemplateTextStreamingModel<INPUT_PROMPT, string, LlamaCppTextGenerationModelSettings<CONTEXT_WINDOW_SIZE>, PromptTemplateTextStreamingModel<string, LlamaCppTextGenerationPrompt, LlamaCppTextGenerationModelSettings<CONTEXT_WINDOW_SIZE>, this>>;
+    withTextPromptTemplate<INPUT_PROMPT>(promptTemplate: TextGenerationPromptTemplate<INPUT_PROMPT, string>): PromptTemplateTextStreamingModel<INPUT_PROMPT, string, LlamaCppCompletionModelSettings<CONTEXT_WINDOW_SIZE>, PromptTemplateTextStreamingModel<string, LlamaCppCompletionPrompt, LlamaCppCompletionModelSettings<CONTEXT_WINDOW_SIZE>, this>>;
     /**
      * Maps the prompt for the full Llama.cpp prompt template (incl. image support).
      */
-    withPromptTemplate<INPUT_PROMPT>(promptTemplate: TextGenerationPromptTemplate<INPUT_PROMPT, LlamaCppTextGenerationPrompt>): PromptTemplateTextStreamingModel<INPUT_PROMPT, LlamaCppTextGenerationPrompt, LlamaCppTextGenerationModelSettings<CONTEXT_WINDOW_SIZE>, this>;
-    withSettings(additionalSettings: Partial<LlamaCppTextGenerationModelSettings<CONTEXT_WINDOW_SIZE>>): this;
+    withPromptTemplate<INPUT_PROMPT>(promptTemplate: TextGenerationPromptTemplate<INPUT_PROMPT, LlamaCppCompletionPrompt>): PromptTemplateTextStreamingModel<INPUT_PROMPT, LlamaCppCompletionPrompt, LlamaCppCompletionModelSettings<CONTEXT_WINDOW_SIZE>, this>;
+    withSettings(additionalSettings: Partial<LlamaCppCompletionModelSettings<CONTEXT_WINDOW_SIZE>>): this;
 }
 declare const llamaCppTextGenerationResponseSchema: z.ZodObject<{
     content: z.ZodString;
@@ -202,7 +289,7 @@ declare const llamaCppTextGenerationResponseSchema: z.ZodObject<{
         seed: z.ZodNumber;
         stop: z.ZodArray<z.ZodString, "many">;
         stream: z.ZodBoolean;
-        temp: z.ZodNumber;
+        temperature: z.ZodOptional<z.ZodNumber>;
         tfs_z: z.ZodNumber;
         top_k: z.ZodNumber;
         top_p: z.ZodNumber;
@@ -226,11 +313,11 @@ declare const llamaCppTextGenerationResponseSchema: z.ZodObject<{
         presence_penalty: number;
         repeat_last_n: number;
         repeat_penalty: number;
-        temp: number;
         tfs_z: number;
         top_k: number;
         top_p: number;
         typical_p: number;
+        temperature?: number | undefined;
     }, {
         model: string;
         stream: boolean;
@@ -250,11 +337,11 @@ declare const llamaCppTextGenerationResponseSchema: z.ZodObject<{
         presence_penalty: number;
         repeat_last_n: number;
         repeat_penalty: number;
-        temp: number;
         tfs_z: number;
         top_k: number;
         top_p: number;
         typical_p: number;
+        temperature?: number | undefined;
     }>;
     model: z.ZodString;
     prompt: z.ZodString;
@@ -267,7 +354,7 @@ declare const llamaCppTextGenerationResponseSchema: z.ZodObject<{
         predicted_n: z.ZodNumber;
         predicted_per_second: z.ZodNullable<z.ZodNumber>;
         predicted_per_token_ms: z.ZodNullable<z.ZodNumber>;
-        prompt_ms: z.ZodNullable<z.ZodNumber>;
+        prompt_ms: z.ZodOptional<z.ZodNullable<z.ZodNumber>>;
         prompt_n: z.ZodNumber;
         prompt_per_second: z.ZodNullable<z.ZodNumber>;
         prompt_per_token_ms: z.ZodNullable<z.ZodNumber>;
@@ -276,19 +363,19 @@ declare const llamaCppTextGenerationResponseSchema: z.ZodObject<{
         predicted_n: number;
         predicted_per_second: number | null;
         predicted_per_token_ms: number | null;
-        prompt_ms: number | null;
         prompt_n: number;
         prompt_per_second: number | null;
         prompt_per_token_ms: number | null;
+        prompt_ms?: number | null | undefined;
     }, {
         predicted_ms: number;
         predicted_n: number;
         predicted_per_second: number | null;
         predicted_per_token_ms: number | null;
-        prompt_ms: number | null;
         prompt_n: number;
         prompt_per_second: number | null;
         prompt_per_token_ms: number | null;
+        prompt_ms?: number | null | undefined;
     }>;
     tokens_cached: z.ZodNumber;
     tokens_evaluated: z.ZodNumber;
@@ -318,11 +405,11 @@ declare const llamaCppTextGenerationResponseSchema: z.ZodObject<{
         presence_penalty: number;
         repeat_last_n: number;
         repeat_penalty: number;
-        temp: number;
         tfs_z: number;
         top_k: number;
         top_p: number;
         typical_p: number;
+        temperature?: number | undefined;
     };
     stopped_eos: boolean;
     stopped_limit: boolean;
@@ -333,10 +420,10 @@ declare const llamaCppTextGenerationResponseSchema: z.ZodObject<{
         predicted_n: number;
         predicted_per_second: number | null;
         predicted_per_token_ms: number | null;
-        prompt_ms: number | null;
         prompt_n: number;
         prompt_per_second: number | null;
         prompt_per_token_ms: number | null;
+        prompt_ms?: number | null | undefined;
     };
     tokens_cached: number;
     tokens_evaluated: number;
@@ -366,11 +453,11 @@ declare const llamaCppTextGenerationResponseSchema: z.ZodObject<{
         presence_penalty: number;
         repeat_last_n: number;
         repeat_penalty: number;
-        temp: number;
         tfs_z: number;
         top_k: number;
         top_p: number;
         typical_p: number;
+        temperature?: number | undefined;
     };
     stopped_eos: boolean;
     stopped_limit: boolean;
@@ -381,10 +468,10 @@ declare const llamaCppTextGenerationResponseSchema: z.ZodObject<{
         predicted_n: number;
         predicted_per_second: number | null;
         predicted_per_token_ms: number | null;
-        prompt_ms: number | null;
         prompt_n: number;
         prompt_per_second: number | null;
         prompt_per_token_ms: number | null;
+        prompt_ms?: number | null | undefined;
     };
     tokens_cached: number;
     tokens_evaluated: number;
@@ -416,11 +503,11 @@ declare const llamaCppTextStreamChunkSchema: import("../../core/schema/ZodSchema
         presence_penalty: number;
         repeat_last_n: number;
         repeat_penalty: number;
-        temp: number;
         tfs_z: number;
         top_k: number;
         top_p: number;
         typical_p: number;
+        temperature?: number | undefined;
     };
     stopped_eos: boolean;
     stopped_limit: boolean;
@@ -431,10 +518,10 @@ declare const llamaCppTextStreamChunkSchema: import("../../core/schema/ZodSchema
         predicted_n: number;
         predicted_per_second: number | null;
         predicted_per_token_ms: number | null;
-        prompt_ms: number | null;
         prompt_n: number;
         prompt_per_second: number | null;
         prompt_per_token_ms: number | null;
+        prompt_ms?: number | null | undefined;
     };
     tokens_cached: number;
     tokens_evaluated: number;
@@ -445,11 +532,11 @@ declare const llamaCppTextStreamChunkSchema: import("../../core/schema/ZodSchema
     content: string;
 }>;
 export type LlamaCppTextStreamChunk = (typeof llamaCppTextStreamChunkSchema)["_type"];
-export type LlamaCppTextGenerationResponseFormatType<T> = {
+export type LlamaCppCompletionResponseFormatType<T> = {
     stream: boolean;
     handler: ResponseHandler<T>;
 };
-export declare const LlamaCppTextGenerationResponseFormat: {
+export declare const LlamaCppCompletionResponseFormat: {
     /**
      * Returns the response as a JSON object.
      */
@@ -479,11 +566,11 @@ export declare const LlamaCppTextGenerationResponseFormat: {
                 presence_penalty: number;
                 repeat_last_n: number;
                 repeat_penalty: number;
-                temp: number;
                 tfs_z: number;
                 top_k: number;
                 top_p: number;
                 typical_p: number;
+                temperature?: number | undefined;
             };
             stopped_eos: boolean;
             stopped_limit: boolean;
@@ -494,10 +581,10 @@ export declare const LlamaCppTextGenerationResponseFormat: {
                 predicted_n: number;
                 predicted_per_second: number | null;
                 predicted_per_token_ms: number | null;
-                prompt_ms: number | null;
                 prompt_n: number;
                 prompt_per_second: number | null;
                 prompt_per_token_ms: number | null;
+                prompt_ms?: number | null | undefined;
             };
             tokens_cached: number;
             tokens_evaluated: number;
@@ -537,11 +624,11 @@ export declare const LlamaCppTextGenerationResponseFormat: {
                 presence_penalty: number;
                 repeat_last_n: number;
                 repeat_penalty: number;
-                temp: number;
                 tfs_z: number;
                 top_k: number;
                 top_p: number;
                 typical_p: number;
+                temperature?: number | undefined;
             };
             stopped_eos: boolean;
             stopped_limit: boolean;
@@ -552,10 +639,10 @@ export declare const LlamaCppTextGenerationResponseFormat: {
                 predicted_n: number;
                 predicted_per_second: number | null;
                 predicted_per_token_ms: number | null;
-                prompt_ms: number | null;
                 prompt_n: number;
                 prompt_per_second: number | null;
                 prompt_per_token_ms: number | null;
+                prompt_ms?: number | null | undefined;
             };
             tokens_cached: number;
             tokens_evaluated: number;

package/model-provider/llamacpp/{LlamaCppTextGenerationModel.js → LlamaCppCompletionModel.js}

@@ -11,7 +11,7 @@ import { parseEventSourceStream } from "../../util/streaming/parseEventSourceStr
 import { LlamaCppApiConfiguration } from "./LlamaCppApiConfiguration.js";
 import { failedLlamaCppCallResponseHandler } from "./LlamaCppError.js";
 import { LlamaCppTokenizer } from "./LlamaCppTokenizer.js";
-export class LlamaCppTextGenerationModel extends AbstractModel {
+export class LlamaCppCompletionModel extends AbstractModel {
     constructor(settings = {}) {
         super({ settings });
         Object.defineProperty(this, "provider", {
@@ -53,10 +53,10 @@ export class LlamaCppTextGenerationModel extends AbstractModel {
                             data,
                         }))
                         : undefined,
-                    cache_prompt: this.settings.cachePrompt,
                     temperature: this.settings.temperature,
                     top_k: this.settings.topK,
                     top_p: this.settings.topP,
+                    min_p: this.settings.minP,
                     n_predict: this.settings.maxGenerationTokens,
                     n_keep: this.settings.nKeep,
                     stop: this.settings.stopSequences,
@@ -65,12 +65,19 @@ export class LlamaCppTextGenerationModel extends AbstractModel {
                     repeat_penalty: this.settings.repeatPenalty,
                     repeat_last_n: this.settings.repeatLastN,
                     penalize_nl: this.settings.penalizeNl,
+                    presence_penalty: this.settings.presencePenalty,
+                    frequency_penalty: this.settings.frequencyPenalty,
+                    penalty_prompt: this.settings.penaltyPrompt,
                     mirostat: this.settings.mirostat,
                     mirostat_tau: this.settings.mirostatTau,
                     mirostat_eta: this.settings.mirostatEta,
+                    grammar: this.settings.grammar,
                     seed: this.settings.seed,
                     ignore_eos: this.settings.ignoreEos,
                     logit_bias: this.settings.logitBias,
+                    n_probs: this.settings.nProbs,
+                    cache_prompt: this.settings.cachePrompt,
+                    slot_id: this.settings.slotId,
                 },
                 failedResponseHandler: failedLlamaCppCallResponseHandler,
                 successfulResponseHandler: responseFormat.handler,
@@ -82,22 +89,29 @@ export class LlamaCppTextGenerationModel extends AbstractModel {
         const eventSettingProperties = [
             ...textGenerationModelProperties,
             "contextWindowSize",
-            "cachePrompt",
             "temperature",
             "topK",
             "topP",
+            "minP",
             "nKeep",
             "tfsZ",
             "typicalP",
             "repeatPenalty",
             "repeatLastN",
             "penalizeNl",
+            "presencePenalty",
+            "frequencyPenalty",
+            "penaltyPrompt",
             "mirostat",
             "mirostatTau",
             "mirostatEta",
+            "grammar",
             "seed",
             "ignoreEos",
             "logitBias",
+            "nProbs",
+            "cachePrompt",
+            "slotId",
         ];
         return Object.fromEntries(Object.entries(this.settings).filter(([key]) => eventSettingProperties.includes(key)));
     }
@@ -108,7 +122,7 @@ export class LlamaCppTextGenerationModel extends AbstractModel {
     async doGenerateTexts(prompt, options) {
         const response = await this.callAPI(prompt, {
             ...options,
-            responseFormat: LlamaCppTextGenerationResponseFormat.json,
+            responseFormat: LlamaCppCompletionResponseFormat.json,
         });
         return {
             response,
@@ -132,7 +146,7 @@ export class LlamaCppTextGenerationModel extends AbstractModel {
     doStreamText(prompt, options) {
         return this.callAPI(prompt, {
             ...options,
-            responseFormat: LlamaCppTextGenerationResponseFormat.deltaIterable,
+            responseFormat: LlamaCppCompletionResponseFormat.deltaIterable,
         });
     }
     extractTextDelta(delta) {
@@ -175,7 +189,7 @@ export class LlamaCppTextGenerationModel extends AbstractModel {
         });
     }
     withSettings(additionalSettings) {
-        return new LlamaCppTextGenerationModel(Object.assign({}, this.settings, additionalSettings));
+        return new LlamaCppCompletionModel(Object.assign({}, this.settings, additionalSettings));
     }
 }
 const llamaCppTextGenerationResponseSchema = z.object({
@@ -200,7 +214,7 @@ const llamaCppTextGenerationResponseSchema = z.object({
         seed: z.number(),
         stop: z.array(z.string()),
         stream: z.boolean(),
-        temp: z.number(),
+        temperature: z.number().optional(), // optional for backwards compatibility
         tfs_z: z.number(),
         top_k: z.number(),
         top_p: z.number(),
@@ -217,7 +231,7 @@ const llamaCppTextGenerationResponseSchema = z.object({
         predicted_n: z.number(),
         predicted_per_second: z.number().nullable(),
         predicted_per_token_ms: z.number().nullable(),
-        prompt_ms: z.number().nullable(),
+        prompt_ms: z.number().nullable().optional(),
         prompt_n: z.number(),
         prompt_per_second: z.number().nullable(),
         prompt_per_token_ms: z.number().nullable(),
@@ -263,7 +277,7 @@ async function createLlamaCppFullDeltaIterableQueue(stream) {
     });
     return queue;
 }
-export const LlamaCppTextGenerationResponseFormat = {
+export const LlamaCppCompletionResponseFormat = {
     /**
      * Returns the response as a JSON object.
      */

package/model-provider/llamacpp/{LlamaCppTextGenerationModel.test.cjs → LlamaCppCompletionModel.test.cjs}

@@ -3,7 +3,7 @@ Object.defineProperty(exports, "__esModule", { value: true });
 const streamText_js_1 = require("../../model-function/generate-text/streamText.cjs");
 const StreamingTestServer_js_1 = require("../../test/StreamingTestServer.cjs");
 const arrayFromAsync_js_1 = require("../../test/arrayFromAsync.cjs");
-const LlamaCppTextGenerationModel_js_1 = require("./LlamaCppTextGenerationModel.cjs");
+const LlamaCppCompletionModel_js_1 = require("./LlamaCppCompletionModel.cjs");
 describe("streamText", () => {
     const server = new StreamingTestServer_js_1.StreamingTestServer("http://127.0.0.1:8080/completion");
     server.setupTestEnvironment();
@@ -17,7 +17,7 @@ describe("streamText", () => {
                 `"mirostat_eta":0.10000000149011612,"mirostat_tau":5.0,"model":"models/llama-2-7b-chat.Q4_K_M.gguf",` +
                 `"n_ctx":4096,"n_keep":0,"n_predict":-1,"n_probs":0,"penalize_nl":true,"penalty_prompt_tokens":[],` +
                 `"presence_penalty":0.0,"repeat_last_n":64,"repeat_penalty":1.100000023841858,"seed":4294967295,` +
-                `"stop":[],"stream":true,"temp":0.800000011920929,"tfs_z":1.0,"top_k":40,"top_p":0.949999988079071,` +
+                `"stop":[],"stream":true,"temperature":0.800000011920929,"tfs_z":1.0,"top_k":40,"top_p":0.949999988079071,` +
                 `"typical_p":1.0,"use_penalty_prompt_tokens":false},"model":"models/llama-2-7b-chat.Q4_K_M.gguf",` +
                 `"prompt":"hello","slot_id":0,"stop":true,"stopped_eos":true,"stopped_limit":false,` +
                 `"stopped_word":false,"stopping_word":"","timings":{"predicted_ms":1054.704,"predicted_n":69,` +
@@ -26,7 +26,7 @@ describe("streamText", () => {
                 `"prompt_per_token_ms":48.845600000000005},"tokens_cached":74,"tokens_evaluated":5,` +
                 `"tokens_predicted":69,"truncated":false}\n\n`,
         ];
-        const stream = await (0, streamText_js_1.streamText)(new LlamaCppTextGenerationModel_js_1.LlamaCppTextGenerationModel().withTextPrompt(), "hello");
+        const stream = await (0, streamText_js_1.streamText)(new LlamaCppCompletionModel_js_1.LlamaCppCompletionModel().withTextPrompt(), "hello");
         // note: space moved to last chunk bc of trimming
         expect(await (0, arrayFromAsync_js_1.arrayFromAsync)(stream)).toStrictEqual([
             "Hello",

package/model-provider/llamacpp/LlamaCppCompletionModel.test.d.ts

@@ -0,0 +1 @@
+export {};

package/model-provider/llamacpp/{LlamaCppTextGenerationModel.test.js → LlamaCppCompletionModel.test.js}

@@ -1,7 +1,7 @@
 import { streamText } from "../../model-function/generate-text/streamText.js";
 import { StreamingTestServer } from "../../test/StreamingTestServer.js";
 import { arrayFromAsync } from "../../test/arrayFromAsync.js";
-import { LlamaCppTextGenerationModel } from "./LlamaCppTextGenerationModel.js";
+import { LlamaCppCompletionModel } from "./LlamaCppCompletionModel.js";
 describe("streamText", () => {
     const server = new StreamingTestServer("http://127.0.0.1:8080/completion");
     server.setupTestEnvironment();
@@ -15,7 +15,7 @@ describe("streamText", () => {
                 `"mirostat_eta":0.10000000149011612,"mirostat_tau":5.0,"model":"models/llama-2-7b-chat.Q4_K_M.gguf",` +
                 `"n_ctx":4096,"n_keep":0,"n_predict":-1,"n_probs":0,"penalize_nl":true,"penalty_prompt_tokens":[],` +
                 `"presence_penalty":0.0,"repeat_last_n":64,"repeat_penalty":1.100000023841858,"seed":4294967295,` +
-                `"stop":[],"stream":true,"temp":0.800000011920929,"tfs_z":1.0,"top_k":40,"top_p":0.949999988079071,` +
+                `"stop":[],"stream":true,"temperature":0.800000011920929,"tfs_z":1.0,"top_k":40,"top_p":0.949999988079071,` +
                 `"typical_p":1.0,"use_penalty_prompt_tokens":false},"model":"models/llama-2-7b-chat.Q4_K_M.gguf",` +
                 `"prompt":"hello","slot_id":0,"stop":true,"stopped_eos":true,"stopped_limit":false,` +
                 `"stopped_word":false,"stopping_word":"","timings":{"predicted_ms":1054.704,"predicted_n":69,` +
@@ -24,7 +24,7 @@ describe("streamText", () => {
                 `"prompt_per_token_ms":48.845600000000005},"tokens_cached":74,"tokens_evaluated":5,` +
                 `"tokens_predicted":69,"truncated":false}\n\n`,
         ];
-        const stream = await streamText(new LlamaCppTextGenerationModel().withTextPrompt(), "hello");
+        const stream = await streamText(new LlamaCppCompletionModel().withTextPrompt(), "hello");
         // note: space moved to last chunk bc of trimming
         expect(await arrayFromAsync(stream)).toStrictEqual([
             "Hello",

package/model-provider/llamacpp/LlamaCppFacade.cjs

@@ -3,10 +3,10 @@ Object.defineProperty(exports, "__esModule", { value: true });
 exports.Tokenizer = exports.TextEmbedder = exports.TextGenerator = void 0;
 const LlamaCppApiConfiguration_js_1 = require("./LlamaCppApiConfiguration.cjs");
 const LlamaCppTextEmbeddingModel_js_1 = require("./LlamaCppTextEmbeddingModel.cjs");
-const LlamaCppTextGenerationModel_js_1 = require("./LlamaCppTextGenerationModel.cjs");
+const LlamaCppCompletionModel_js_1 = require("./LlamaCppCompletionModel.cjs");
 const LlamaCppTokenizer_js_1 = require("./LlamaCppTokenizer.cjs");
 function TextGenerator(settings = {}) {
-    return new LlamaCppTextGenerationModel_js_1.LlamaCppTextGenerationModel(settings);
+    return new LlamaCppCompletionModel_js_1.LlamaCppCompletionModel(settings);
 }
 exports.TextGenerator = TextGenerator;
 function TextEmbedder(settings = {}) {

package/model-provider/llamacpp/LlamaCppFacade.d.ts

@@ -1,7 +1,7 @@
 import { ApiConfiguration } from "../../core/api/ApiConfiguration.js";
 import { LlamaCppTextEmbeddingModel, LlamaCppTextEmbeddingModelSettings } from "./LlamaCppTextEmbeddingModel.js";
-import { LlamaCppTextGenerationModel, LlamaCppTextGenerationModelSettings } from "./LlamaCppTextGenerationModel.js";
+import { LlamaCppCompletionModel, LlamaCppCompletionModelSettings } from "./LlamaCppCompletionModel.js";
 import { LlamaCppTokenizer } from "./LlamaCppTokenizer.js";
-export declare function TextGenerator<CONTEXT_WINDOW_SIZE extends number>(settings?: LlamaCppTextGenerationModelSettings<CONTEXT_WINDOW_SIZE>): LlamaCppTextGenerationModel<CONTEXT_WINDOW_SIZE>;
+export declare function TextGenerator<CONTEXT_WINDOW_SIZE extends number>(settings?: LlamaCppCompletionModelSettings<CONTEXT_WINDOW_SIZE>): LlamaCppCompletionModel<CONTEXT_WINDOW_SIZE>;
 export declare function TextEmbedder(settings?: LlamaCppTextEmbeddingModelSettings): LlamaCppTextEmbeddingModel;
 export declare function Tokenizer(api?: ApiConfiguration): LlamaCppTokenizer;

package/model-provider/llamacpp/LlamaCppFacade.js

@@ -1,9 +1,9 @@
 import { LlamaCppApiConfiguration } from "./LlamaCppApiConfiguration.js";
 import { LlamaCppTextEmbeddingModel, } from "./LlamaCppTextEmbeddingModel.js";
-import { LlamaCppTextGenerationModel, } from "./LlamaCppTextGenerationModel.js";
+import { LlamaCppCompletionModel, } from "./LlamaCppCompletionModel.js";
 import { LlamaCppTokenizer } from "./LlamaCppTokenizer.js";
 export function TextGenerator(settings = {}) {
-    return new LlamaCppTextGenerationModel(settings);
+    return new LlamaCppCompletionModel(settings);
 }
 export function TextEmbedder(settings = {}) {
     return new LlamaCppTextEmbeddingModel(settings);

package/model-provider/llamacpp/index.cjs

@@ -33,5 +33,5 @@ var LlamaCppError_js_1 = require("./LlamaCppError.cjs");
 Object.defineProperty(exports, "LlamaCppError", { enumerable: true, get: function () { return LlamaCppError_js_1.LlamaCppError; } });
 exports.llamacpp = __importStar(require("./LlamaCppFacade.cjs"));
 __exportStar(require("./LlamaCppTextEmbeddingModel.cjs"), exports);
-__exportStar(require("./LlamaCppTextGenerationModel.cjs"), exports);
+__exportStar(require("./LlamaCppCompletionModel.cjs"), exports);
 __exportStar(require("./LlamaCppTokenizer.cjs"), exports);