final chores & docs

chmjkb · chmjkb · commit a22dd0fea4af · 2026-03-12T16:38:41.000+01:00
diff --git a/docs/docs/04-typescript-api/01-natural-language-processing/LLMModule.md b/docs/docs/04-typescript-api/01-natural-language-processing/LLMModule.md
@@ -129,10 +129,10 @@ The `capabilities` field is already set on the model constant. You can also cons
 
 ```typescript
 const llm = await LLMModule.fromModelName({
-  modelName: 'custom',
-  modelSource: '...',
-  tokenizerSource: '...',
-  tokenizerConfigSource: '...',
+  modelName: 'lfm2.5-vl-1.6b-quantized',
+  modelSource: require('./path/to/model.pte'),
+  tokenizerSource: require('./path/to/tokenizer.json'),
+  tokenizerConfigSource: require('./path/to/tokenizer_config.json'),
   capabilities: ['vision'],
 });
 ```
diff --git a/docs/docs/04-typescript-api/01-natural-language-processing/TextToSpeechModule.md b/docs/docs/04-typescript-api/01-natural-language-processing/TextToSpeechModule.md
@@ -39,7 +39,7 @@ Use the static [`fromModelName`](../../06-api-reference/classes/TextToSpeechModu
   - [`model`](../../06-api-reference/interfaces/TextToSpeechConfig.md#model) - Model configuration (e.g. `KOKORO_MEDIUM`).
   - [`voice`](../../06-api-reference/interfaces/TextToSpeechConfig.md#voice) - Voice configuration (e.g. `KOKORO_VOICE_AF_HEART`).
 
-- `onDownloadProgress` - Optional callback to track download progress (value between 0 and 1).
+- [`onDownloadProgress`](../../06-api-reference/classes/TextToSpeechModule.md#frommodelname) - Optional callback to track download progress (value between 0 and 1).
 
 This method returns a promise that resolves to a `TextToSpeechModule` instance once the assets are downloaded and loaded into memory.
 
diff --git a/docs/docs/04-typescript-api/01-natural-language-processing/VADModule.md b/docs/docs/04-typescript-api/01-natural-language-processing/VADModule.md
@@ -14,9 +14,8 @@ TypeScript API implementation of the [useVAD](../../03-hooks/01-natural-language
 ```typescript
 import { VADModule, FSMN_VAD } from 'react-native-executorch';
 
-const model = await VADModule.fromModelName(
-  { modelSource: FSMN_VAD },
-  (progress) => console.log(progress)
+const model = await VADModule.fromModelName(FSMN_VAD, (progress) =>
+  console.log(progress)
 );
 
 await model.forward(waveform);
@@ -30,7 +29,8 @@ All methods of `VADModule` are explained in details here: [`VADModule` API Refer
 
 To create a ready-to-use instance, call the static [`fromModelName`](../../06-api-reference/classes/VADModule.md#frommodelname) factory with the following parameters:
 
-- `model` - Object containing:
+- `namedSources` - Object containing:
+  - `modelName` - Model name identifier.
   - `modelSource` - Location of the model binary.
 
 - `onDownloadProgress` - Optional callback to track download progress (value between 0 and 1).
diff --git a/docs/docs/04-typescript-api/02-computer-vision/ClassificationModule.md b/docs/docs/04-typescript-api/02-computer-vision/ClassificationModule.md
@@ -20,9 +20,8 @@ import {
 const imageUri = 'path/to/image.png';
 
 // Creating and loading the module
-const classificationModule = await ClassificationModule.fromModelName({
-  modelSource: EFFICIENTNET_V2_S,
-});
+const classificationModule =
+  await ClassificationModule.fromModelName(EFFICIENTNET_V2_S);
 
 // Running the model
 const classesWithProbabilities = await classificationModule.forward(imageUri);
@@ -36,7 +35,8 @@ All methods of `ClassificationModule` are explained in details here: [`Classific
 
 To create a ready-to-use instance, call the static [`fromModelName`](../../06-api-reference/classes/ClassificationModule.md#frommodelname) factory with the following parameters:
 
-- `model` - Object containing:
+- `namedSources` - Object containing:
+  - `modelName` - Model name identifier.
   - `modelSource` - Location of the model binary.
 
 - `onDownloadProgress` - Optional callback to track download progress (value between 0 and 1).
diff --git a/docs/docs/04-typescript-api/02-computer-vision/ImageEmbeddingsModule.md b/docs/docs/04-typescript-api/02-computer-vision/ImageEmbeddingsModule.md
@@ -18,9 +18,9 @@ import {
 } from 'react-native-executorch';
 
 // Creating and loading the module
-const imageEmbeddingsModule = await ImageEmbeddingsModule.fromModelName({
-  modelSource: CLIP_VIT_BASE_PATCH32_IMAGE,
-});
+const imageEmbeddingsModule = await ImageEmbeddingsModule.fromModelName(
+  CLIP_VIT_BASE_PATCH32_IMAGE
+);
 
 // Running the model
 const embedding = await imageEmbeddingsModule.forward(
@@ -36,7 +36,8 @@ All methods of `ImageEmbeddingsModule` are explained in details here: [`ImageEmb
 
 To create a ready-to-use instance, call the static [`fromModelName`](../../06-api-reference/classes/ImageEmbeddingsModule.md#frommodelname) factory with the following parameters:
 
-- `model` - Object containing:
+- `namedSources` - Object containing:
+  - `modelName` - Model name identifier.
   - `modelSource` - Location of the model binary.
 
 - `onDownloadProgress` - Optional callback to track download progress (value between 0 and 1).
diff --git a/docs/docs/04-typescript-api/02-computer-vision/OCRModule.md b/docs/docs/04-typescript-api/02-computer-vision/OCRModule.md
@@ -28,8 +28,9 @@ All methods of `OCRModule` are explained in details here: [`OCRModule` API Refer
 
 ## Loading the model
 
-Use the static [`fromModelName`](../../06-api-reference/classes/OCRModule.md#frommodelname) factory method. It accepts a model config object (e.g. `OCR_ENGLISH`) containing:
+Use the static [`fromModelName`](../../06-api-reference/classes/OCRModule.md#frommodelname) factory method. It accepts a `namedSources` object (e.g. `OCR_ENGLISH`) containing:
 
+- `modelName` - Model name identifier.
 - [`detectorSource`](../../06-api-reference/classes/OCRModule.md#detectorsource) - Location of the used detector.
 - [`recognizerSource`](../../06-api-reference/classes/OCRModule.md#recognizersource) - Location of the used recognizer.
 - [`language`](../../06-api-reference/classes/OCRModule.md#recognizersource) - Language used in OCR.
diff --git a/docs/docs/04-typescript-api/02-computer-vision/StyleTransferModule.md b/docs/docs/04-typescript-api/02-computer-vision/StyleTransferModule.md
@@ -20,9 +20,8 @@ import {
 const imageUri = 'path/to/image.png';
 
 // Creating and loading the module
-const styleTransferModule = await StyleTransferModule.fromModelName({
-  modelSource: STYLE_TRANSFER_CANDY,
-});
+const styleTransferModule =
+  await StyleTransferModule.fromModelName(STYLE_TRANSFER_CANDY);
 
 // Running the model
 const generatedImageUrl = await styleTransferModule.forward(imageUri);
@@ -36,7 +35,8 @@ All methods of `StyleTransferModule` are explained in details here: [`StyleTrans
 
 To create a ready-to-use instance, call the static [`fromModelName`](../../06-api-reference/classes/StyleTransferModule.md#frommodelname) factory with the following parameters:
 
-- `model` - Object containing:
+- `namedSources` - Object containing:
+  - `modelName` - Model name identifier.
   - `modelSource` - Location of the model binary.
 
 - `onDownloadProgress` - Optional callback to track download progress (value between 0 and 1).
diff --git a/docs/docs/04-typescript-api/02-computer-vision/VerticalOCRModule.md b/docs/docs/04-typescript-api/02-computer-vision/VerticalOCRModule.md
@@ -29,8 +29,9 @@ All methods of `VerticalOCRModule` are explained in details here: [`VerticalOCRM
 
 ## Loading the model
 
-Use the static [`fromModelName`](../../06-api-reference/classes/VerticalOCRModule.md#frommodelname) factory method. It accepts a model config object (e.g. `{ ...OCR_ENGLISH, independentCharacters: true }`) containing:
+Use the static [`fromModelName`](../../06-api-reference/classes/VerticalOCRModule.md#frommodelname) factory method. It accepts a `namedSources` object (e.g. `{ ...OCR_ENGLISH, independentCharacters: true }`) containing:
 
+- `modelName` - Model name identifier.
 - [`detectorSource`](../../06-api-reference/classes/VerticalOCRModule.md#detectorsource) - Location of the used detector.
 - [`recognizerSource`](../../06-api-reference/classes/VerticalOCRModule.md#recognizersource) - Location of the used recognizer.
 - [`language`](../../06-api-reference/classes/VerticalOCRModule.md#recognizersource) - Language used in OCR.
diff --git a/packages/react-native-executorch/src/modules/computer_vision/OCRModule.ts b/packages/react-native-executorch/src/modules/computer_vision/OCRModule.ts
@@ -12,8 +12,8 @@ import { parseUnknownError } from '../../errors/errorUtils';
 export class OCRModule {
   private controller: OCRController;
 
-  private constructor() {
-    this.controller = new OCRController();
+  private constructor(controller: OCRController) {
+    this.controller = controller;
   }
 
   /**
@@ -38,15 +38,15 @@ export class OCRModule {
     },
     onDownloadProgress: (progress: number) => void = () => {}
   ): Promise<OCRModule> {
-    const instance = new OCRModule();
     try {
-      await instance.controller.load(
+      const controller = new OCRController();
+      await controller.load(
         namedSources.detectorSource,
         namedSources.recognizerSource,
         namedSources.language,
         onDownloadProgress
       );
-      return instance;
+      return new OCRModule(controller);
     } catch (error) {
       Logger.error('Load failed:', error);
       throw parseUnknownError(error);
diff --git a/packages/react-native-executorch/src/modules/computer_vision/VerticalOCRModule.ts b/packages/react-native-executorch/src/modules/computer_vision/VerticalOCRModule.ts
@@ -12,8 +12,8 @@ import { OCRDetection, OCRLanguage, OCRModelName } from '../../types/ocr';
 export class VerticalOCRModule {
   private controller: VerticalOCRController;
 
-  private constructor() {
-    this.controller = new VerticalOCRController();
+  private constructor(controller: VerticalOCRController) {
+    this.controller = controller;
   }
 
   /**
@@ -39,16 +39,16 @@ export class VerticalOCRModule {
     },
     onDownloadProgress: (progress: number) => void = () => {}
   ): Promise<VerticalOCRModule> {
-    const instance = new VerticalOCRModule();
     try {
-      await instance.controller.load(
+      const controller = new VerticalOCRController();
+      await controller.load(
         namedSources.detectorSource,
         namedSources.recognizerSource,
         namedSources.language,
         namedSources.independentCharacters ?? false,
         onDownloadProgress
       );
-      return instance;
+      return new VerticalOCRModule(controller);
     } catch (error) {
       Logger.error('Load failed:', error);
       throw parseUnknownError(error);
diff --git a/packages/react-native-executorch/src/modules/natural_language_processing/SpeechToTextModule.ts b/packages/react-native-executorch/src/modules/natural_language_processing/SpeechToTextModule.ts
@@ -17,9 +17,15 @@ import { Logger } from '../../common/Logger';
  */
 export class SpeechToTextModule {
   private nativeModule: any;
-  private modelConfig!: SpeechToTextModelConfig;
+  private modelConfig: SpeechToTextModelConfig;
 
-  private constructor() {}
+  private constructor(
+    nativeModule: unknown,
+    modelConfig: SpeechToTextModelConfig
+  ) {
+    this.nativeModule = nativeModule;
+    this.modelConfig = modelConfig;
+  }
 
   /**
    * Creates a Speech to Text instance for a built-in model.
@@ -38,10 +44,12 @@ export class SpeechToTextModule {
     namedSources: SpeechToTextModelConfig,
     onDownloadProgress: (progress: number) => void = () => {}
   ): Promise<SpeechToTextModule> {
-    const instance = new SpeechToTextModule();
     try {
-      await instance.internalLoad(namedSources, onDownloadProgress);
-      return instance;
+      const nativeModule = await SpeechToTextModule.loadWhisper(
+        namedSources,
+        onDownloadProgress
+      );
+      return new SpeechToTextModule(nativeModule, namedSources);
     } catch (error) {
       Logger.error('Load failed:', error);
       throw parseUnknownError(error);
@@ -80,12 +88,10 @@ export class SpeechToTextModule {
     );
   }
 
-  private async internalLoad(
+  private static async loadWhisper(
     model: SpeechToTextModelConfig,
-    onDownloadProgressCallback: (progress: number) => void = () => {}
-  ) {
-    this.modelConfig = model;
-
+    onDownloadProgressCallback: (progress: number) => void
+  ): Promise<unknown> {
     const tokenizerLoadPromise = ResourceFetcher.fetch(
       undefined,
       model.tokenizerSource
@@ -105,7 +111,7 @@ export class SpeechToTextModule {
       );
     }
     // Currently only Whisper architecture is supported
-    this.nativeModule = await global.loadSpeechToText(
+    return await global.loadSpeechToText(
       'whisper',
       modelSources[0],
       tokenizerSources[0]
diff --git a/packages/react-native-executorch/src/modules/natural_language_processing/TextToSpeechModule.ts b/packages/react-native-executorch/src/modules/natural_language_processing/TextToSpeechModule.ts
@@ -80,7 +80,7 @@ export class TextToSpeechModule {
       voice.extra.lexiconSource
     );
 
-    if (paths === null || paths.length !== 5 || paths.some((p) => p == null)) {
+    if (paths === null || paths.length !== 5) {
       throw new RnExecutorchError(
         RnExecutorchErrorCode.DownloadInterrupted,
         'Download interrupted or missing resource.'
@@ -122,7 +122,7 @@ export class TextToSpeechModule {
     speed: number = 1.0
   ): Promise<Float32Array> {
     this.ensureLoaded('forward');
-    return await (this.nativeModule as any).generate(text, speed);
+    return await this.nativeModule.generate(text, speed);
   }
 
   /**
@@ -139,10 +139,7 @@ export class TextToSpeechModule {
     speed: number = 1.0
   ): Promise<Float32Array> {
     this.ensureLoaded('forwardFromPhonemes');
-    return await (this.nativeModule as any).generateFromPhonemes(
-      phonemes,
-      speed
-    );
+    return await this.nativeModule.generateFromPhonemes(phonemes, speed);
   }
 
   /**
