react-native-executorch 0.4.9 → 0.5.0-nightly-6f4dd53-20251211

package/ios/RnExecutorch/models/ocr/utils/OCRUtils.h

@@ -1,10 +0,0 @@
-#import <opencv2/opencv.hpp>
-
-@interface OCRUtils : NSObject
-
-+ (cv::Mat)resizeWithPadding:(cv::Mat)img
-                desiredWidth:(int)desiredWidth
-               desiredHeight:(int)desiredHeight;
-+ (cv::Rect)extractBoundingBox:(NSArray *)coords;
-
-@end

package/ios/RnExecutorch/models/ocr/utils/OCRUtils.mm

@@ -1,67 +0,0 @@
-#import "OCRUtils.h"
-
-@implementation OCRUtils
-
-+ (cv::Mat)resizeWithPadding:(cv::Mat)img
-                desiredWidth:(int)desiredWidth
-               desiredHeight:(int)desiredHeight {
-  const int height = img.rows;
-  const int width = img.cols;
-  const float heightRatio = (float)desiredHeight / height;
-  const float widthRatio = (float)desiredWidth / width;
-  const float resizeRatio = MIN(heightRatio, widthRatio);
-
-  const int newWidth = width * resizeRatio;
-  const int newHeight = height * resizeRatio;
-
-  cv::Mat resizedImg;
-  cv::resize(img, resizedImg, cv::Size(newWidth, newHeight), 0, 0,
-             cv::INTER_AREA);
-
-  const int cornerPatchSize = MAX(1, MIN(height, width) / 30);
-  std::vector<cv::Mat> corners = {
-      img(cv::Rect(0, 0, cornerPatchSize, cornerPatchSize)),
-      img(cv::Rect(width - cornerPatchSize, 0, cornerPatchSize,
-                   cornerPatchSize)),
-      img(cv::Rect(0, height - cornerPatchSize, cornerPatchSize,
-                   cornerPatchSize)),
-      img(cv::Rect(width - cornerPatchSize, height - cornerPatchSize,
-                   cornerPatchSize, cornerPatchSize))};
-
-  cv::Scalar backgroundScalar = cv::mean(corners[0]);
-  for (int i = 1; i < corners.size(); i++) {
-    backgroundScalar += cv::mean(corners[i]);
-  }
-  backgroundScalar /= (double)corners.size();
-
-  backgroundScalar[0] = cvFloor(backgroundScalar[0]);
-  backgroundScalar[1] = cvFloor(backgroundScalar[1]);
-  backgroundScalar[2] = cvFloor(backgroundScalar[2]);
-
-  const int deltaW = desiredWidth - newWidth;
-  const int deltaH = desiredHeight - newHeight;
-  const int top = deltaH / 2;
-  const int bottom = deltaH - top;
-  const int left = deltaW / 2;
-  const int right = deltaW - left;
-
-  cv::Mat centeredImg;
-  cv::copyMakeBorder(resizedImg, centeredImg, top, bottom, left, right,
-                     cv::BORDER_CONSTANT, backgroundScalar);
-
-  return centeredImg;
-}
-
-+ (cv::Rect)extractBoundingBox:(NSArray *)coords {
-  std::vector<cv::Point2f> points;
-  points.reserve(coords.count);
-  for (NSValue *value in coords) {
-    const CGPoint point = [value CGPointValue];
-
-    points.emplace_back(point.x, point.y);
-  }
-
-  return cv::boundingRect(points);
-}
-
-@end

package/ios/RnExecutorch/models/ocr/utils/RecognizerUtils.h

@@ -1,35 +0,0 @@
-#import <opencv2/opencv.hpp>
-
-@interface RecognizerUtils : NSObject
-
-+ (CGFloat)calculateRatio:(int)width height:(int)height;
-+ (cv::Mat)computeRatioAndResize:(cv::Mat)img
-                           width:(int)width
-                          height:(int)height
-                     modelHeight:(int)modelHeight;
-+ (cv::Mat)normalizeForRecognizer:(cv::Mat)image
-                   adjustContrast:(double)adjustContrast
-                       isVertical:(BOOL)isVertical;
-+ (cv::Mat)adjustContrastGrey:(cv::Mat)img target:(double)target;
-+ (cv::Mat)divideMatrix:(cv::Mat)matrix byVector:(NSArray<NSNumber *> *)vector;
-+ (cv::Mat)softmax:(cv::Mat)inputs;
-+ (NSDictionary *)calculateResizeRatioAndPaddings:(int)width
-                                           height:(int)height
-                                     desiredWidth:(int)desiredWidth
-                                    desiredHeight:(int)desiredHeight;
-+ (cv::Mat)getCroppedImage:(NSDictionary *)box
-                     image:(cv::Mat)image
-               modelHeight:(int)modelHeight;
-+ (NSMutableArray *)sumProbabilityRows:(cv::Mat)probabilities
-                     modelOutputHeight:(int)modelOutputHeight;
-+ (NSArray *)findMaxValuesAndIndices:(cv::Mat)probabilities;
-+ (double)computeConfidenceScore:(NSArray<NSNumber *> *)valuesArray
-                    indicesArray:(NSArray<NSNumber *> *)indicesArray;
-+ (cv::Mat)cropImageWithBoundingBox:(cv::Mat &)img
-                               bbox:(NSArray *)bbox
-                       originalBbox:(NSArray *)originalBbox
-                           paddings:(NSDictionary *)paddings
-                   originalPaddings:(NSDictionary *)originalPaddings;
-+ (cv::Mat)cropSingleCharacter:(cv::Mat)img;
-
-@end

package/ios/RnExecutorch/models/ocr/utils/RecognizerUtils.mm

@@ -1,331 +0,0 @@
-#import "RecognizerUtils.h"
-#import "Constants.h"
-#import "OCRUtils.h"
-
-@implementation RecognizerUtils
-
-+ (CGFloat)calculateRatio:(int)width height:(int)height {
-  CGFloat ratio = (CGFloat)width / (CGFloat)height;
-  if (ratio < 1.0) {
-    ratio = 1.0 / ratio;
-  }
-  return ratio;
-}
-
-+ (cv::Mat)computeRatioAndResize:(cv::Mat)img
-                           width:(int)width
-                          height:(int)height
-                     modelHeight:(int)modelHeight {
-  CGFloat ratio = (CGFloat)width / (CGFloat)height;
-  if (ratio < 1.0) {
-    ratio = [self calculateRatio:width height:height];
-    cv::resize(img, img, cv::Size(modelHeight, (int)(modelHeight * ratio)), 0,
-               0, cv::INTER_LANCZOS4);
-  } else {
-    cv::resize(img, img, cv::Size((int)(modelHeight * ratio), modelHeight), 0,
-               0, cv::INTER_LANCZOS4);
-  }
-  return img;
-}
-
-+ (cv::Mat)adjustContrastGrey:(cv::Mat)img target:(double)target {
-  double contrast = 0.0;
-  int high = 0;
-  int low = 255;
-
-  for (int i = 0; i < img.rows; ++i) {
-    for (int j = 0; j < img.cols; ++j) {
-      uchar pixel = img.at<uchar>(i, j);
-      high = MAX(high, pixel);
-      low = MIN(low, pixel);
-    }
-  }
-  contrast = (high - low) / 255.0;
-
-  if (contrast < target) {
-    const double ratio = 200.0 / MAX(10, high - low);
-    img.convertTo(img, CV_32F);
-    img = ((img - low + 25) * ratio);
-
-    cv::threshold(img, img, 255, 255, cv::THRESH_TRUNC);
-    cv::threshold(img, img, 0, 0, cv::THRESH_TOZERO);
-
-    img.convertTo(img, CV_8U);
-  }
-
-  return img;
-}
-
-+ (cv::Mat)normalizeForRecognizer:(cv::Mat)image
-                   adjustContrast:(double)adjustContrast
-                       isVertical:(BOOL)isVertical {
-  if (adjustContrast > 0) {
-    image = [self adjustContrastGrey:image target:adjustContrast];
-  }
-
-  int desiredWidth =
-      (isVertical) ? smallVerticalRecognizerWidth : smallRecognizerWidth;
-
-  if (image.cols >= largeRecognizerWidth) {
-    desiredWidth = largeRecognizerWidth;
-  } else if (image.cols >= mediumRecognizerWidth) {
-    desiredWidth = mediumRecognizerWidth;
-  }
-
-  image = [OCRUtils resizeWithPadding:image
-                         desiredWidth:desiredWidth
-                        desiredHeight:recognizerHeight];
-
-  image.convertTo(image, CV_32F, 1.0 / 255.0);
-  image = (image - 0.5) * 2.0;
-
-  return image;
-}
-
-+ (cv::Mat)divideMatrix:(cv::Mat)matrix byVector:(NSArray<NSNumber *> *)vector {
-  cv::Mat result = matrix.clone();
-
-  for (int i = 0; i < matrix.rows; i++) {
-    const float divisor = [vector[i] floatValue];
-    for (int j = 0; j < matrix.cols; j++) {
-      result.at<float>(i, j) /= divisor;
-    }
-  }
-
-  return result;
-}
-
-+ (cv::Mat)softmax:(cv::Mat)inputs {
-  cv::Mat maxVal;
-  cv::reduce(inputs, maxVal, 1, cv::REDUCE_MAX, CV_32F);
-  cv::Mat expInputs;
-  cv::exp(inputs - cv::repeat(maxVal, 1, inputs.cols), expInputs);
-  cv::Mat sumExp;
-  cv::reduce(expInputs, sumExp, 1, cv::REDUCE_SUM, CV_32F);
-  cv::Mat softmaxOutput = expInputs / cv::repeat(sumExp, 1, inputs.cols);
-  return softmaxOutput;
-}
-
-+ (NSDictionary *)calculateResizeRatioAndPaddings:(int)width
-                                           height:(int)height
-                                     desiredWidth:(int)desiredWidth
-                                    desiredHeight:(int)desiredHeight {
-  const float newRatioH = (float)desiredHeight / height;
-  const float newRatioW = (float)desiredWidth / width;
-  float resizeRatio = MIN(newRatioH, newRatioW);
-  const int newWidth = width * resizeRatio;
-  const int newHeight = height * resizeRatio;
-  const int deltaW = desiredWidth - newWidth;
-  const int deltaH = desiredHeight - newHeight;
-  const int top = deltaH / 2;
-  const int left = deltaW / 2;
-  const float heightRatio = (float)height / desiredHeight;
-  const float widthRatio = (float)width / desiredWidth;
-
-  resizeRatio = MAX(heightRatio, widthRatio);
-
-  return @{
-    @"resizeRatio" : @(resizeRatio),
-    @"top" : @(top),
-    @"left" : @(left),
-  };
-}
-
-+ (cv::Mat)getCroppedImage:(NSDictionary *)box
-                     image:(cv::Mat)image
-               modelHeight:(int)modelHeight {
-  NSArray *coords = box[@"bbox"];
-  const CGFloat angle = [box[@"angle"] floatValue];
-
-  std::vector<cv::Point2f> points;
-  for (NSValue *value in coords) {
-    const CGPoint point = [value CGPointValue];
-    points.emplace_back(static_cast<float>(point.x),
-                        static_cast<float>(point.y));
-  }
-
-  cv::RotatedRect rotatedRect = cv::minAreaRect(points);
-
-  cv::Point2f imageCenter = cv::Point2f(image.cols / 2.0, image.rows / 2.0);
-  cv::Mat rotationMatrix = cv::getRotationMatrix2D(imageCenter, angle, 1.0);
-  cv::Mat rotatedImage;
-  cv::warpAffine(image, rotatedImage, rotationMatrix, image.size(),
-                 cv::INTER_LINEAR);
-  cv::Point2f rectPoints[4];
-  rotatedRect.points(rectPoints);
-  std::vector<cv::Point2f> transformedPoints(4);
-  cv::Mat rectMat(4, 2, CV_32FC2, rectPoints);
-  cv::transform(rectMat, rectMat, rotationMatrix);
-
-  for (int i = 0; i < 4; ++i) {
-    transformedPoints[i] = rectPoints[i];
-  }
-
-  cv::Rect boundingBox = cv::boundingRect(transformedPoints);
-  boundingBox &= cv::Rect(0, 0, rotatedImage.cols, rotatedImage.rows);
-  cv::Mat croppedImage = rotatedImage(boundingBox);
-  if (boundingBox.width == 0 || boundingBox.height == 0) {
-    croppedImage = cv::Mat().empty();
-
-    return croppedImage;
-  }
-
-  croppedImage = [self computeRatioAndResize:croppedImage
-                                       width:boundingBox.width
-                                      height:boundingBox.height
-                                 modelHeight:modelHeight];
-
-  return croppedImage;
-}
-
-+ (NSMutableArray *)sumProbabilityRows:(cv::Mat)probabilities
-                     modelOutputHeight:(int)modelOutputHeight {
-  NSMutableArray *predsNorm =
-      [NSMutableArray arrayWithCapacity:probabilities.rows];
-  for (int i = 0; i < probabilities.rows; i++) {
-    float sum = 0.0;
-    for (int j = 0; j < modelOutputHeight; j++) {
-      sum += probabilities.at<float>(i, j);
-    }
-    [predsNorm addObject:@(sum)];
-  }
-  return predsNorm;
-}
-
-+ (NSArray *)findMaxValuesAndIndices:(cv::Mat)probabilities {
-  NSMutableArray *valuesArray = [NSMutableArray array];
-  NSMutableArray *indicesArray = [NSMutableArray array];
-  for (int i = 0; i < probabilities.rows; i++) {
-    double maxVal = 0;
-    cv::Point maxLoc;
-    cv::minMaxLoc(probabilities.row(i), NULL, &maxVal, NULL, &maxLoc);
-    [valuesArray addObject:@(maxVal)];
-    [indicesArray addObject:@(maxLoc.x)];
-  }
-  return @[ valuesArray, indicesArray ];
-}
-
-+ (double)computeConfidenceScore:(NSArray<NSNumber *> *)valuesArray
-                    indicesArray:(NSArray<NSNumber *> *)indicesArray {
-  NSMutableArray *predsMaxProb = [NSMutableArray array];
-  for (NSUInteger index = 0; index < indicesArray.count; index++) {
-    NSNumber *indicator = indicesArray[index];
-    if ([indicator intValue] != 0) {
-      [predsMaxProb addObject:valuesArray[index]];
-    }
-  }
-  if (predsMaxProb.count == 0) {
-    [predsMaxProb addObject:@(0)];
-  }
-  double product = 1.0;
-  for (NSNumber *prob in predsMaxProb) {
-    product *= [prob doubleValue];
-  }
-  return pow(product, 2.0 / sqrt(predsMaxProb.count));
-}
-
-+ (cv::Mat)cropImageWithBoundingBox:(cv::Mat &)img
-                               bbox:(NSArray *)bbox
-                       originalBbox:(NSArray *)originalBbox
-                           paddings:(NSDictionary *)paddings
-                   originalPaddings:(NSDictionary *)originalPaddings {
-  CGPoint topLeft = [originalBbox[0] CGPointValue];
-  std::vector<cv::Point2f> points;
-  points.reserve(bbox.count);
-  for (NSValue *coords in bbox) {
-    CGPoint point = [coords CGPointValue];
-
-    point.x = point.x - [paddings[@"left"] intValue];
-    point.y = point.y - [paddings[@"top"] intValue];
-
-    point.x = point.x * [paddings[@"resizeRatio"] floatValue];
-    point.y = point.y * [paddings[@"resizeRatio"] floatValue];
-
-    point.x = point.x + topLeft.x;
-    point.y = point.y + topLeft.y;
-
-    point.x = point.x - [originalPaddings[@"left"] intValue];
-    point.y = point.y - [originalPaddings[@"top"] intValue];
-
-    point.x = point.x * [originalPaddings[@"resizeRatio"] floatValue];
-    point.y = point.y * [originalPaddings[@"resizeRatio"] floatValue];
-
-    points.emplace_back(cv::Point2f(point.x, point.y));
-  }
-
-  cv::Rect rect = cv::boundingRect(points);
-  cv::Mat croppedImage = img(rect);
-  cv::cvtColor(croppedImage, croppedImage, cv::COLOR_BGR2GRAY);
-  cv::resize(croppedImage, croppedImage,
-             cv::Size(smallVerticalRecognizerWidth, recognizerHeight), 0, 0,
-             cv::INTER_AREA);
-  cv::medianBlur(img, img, 1);
-  return croppedImage;
-}
-
-+ (cv::Mat)cropSingleCharacter:(cv::Mat)img {
-
-  cv::Mat histogram;
-
-  int histSize = 256;
-  float range[] = {0, 256};
-  const float *histRange = {range};
-  bool uniform = true, accumulate = false;
-
-  cv::calcHist(&img, 1, 0, cv::Mat(), histogram, 1, &histSize, &histRange,
-               uniform, accumulate);
-
-  int midPoint = histSize / 2;
-
-  double sumLeft = 0.0, sumRight = 0.0;
-  for (int i = 0; i < midPoint; i++) {
-    sumLeft += histogram.at<float>(i);
-  }
-  for (int i = midPoint; i < histSize; i++) {
-    sumRight += histogram.at<float>(i);
-  }
-
-  const int thresholdType =
-      (sumLeft < sumRight) ? cv::THRESH_BINARY_INV : cv::THRESH_BINARY;
-
-  cv::Mat thresh;
-  cv::threshold(img, thresh, 0, 255, thresholdType + cv::THRESH_OTSU);
-
-  cv::Mat labels, stats, centroids;
-  const int numLabels =
-      connectedComponentsWithStats(thresh, labels, stats, centroids, 8);
-  const CGFloat centralThreshold = singleCharacterCenterThreshold;
-  const int height = thresh.rows;
-  const int width = thresh.cols;
-
-  const int minX = centralThreshold * width;
-  const int maxX = (1 - centralThreshold) * width;
-  const int minY = centralThreshold * height;
-  const int maxY = (1 - centralThreshold) * height;
-
-  int selectedComponent = -1;
-
-  for (int i = 1; i < numLabels; i++) {
-    const int area = stats.at<int>(i, cv::CC_STAT_AREA);
-    const double cx = centroids.at<double>(i, 0);
-    const double cy = centroids.at<double>(i, 1);
-
-    if (minX < cx && cx < maxX && minY < cy && cy < maxY &&
-        area > singleCharacterMinSize) {
-      if (selectedComponent == -1 ||
-          area > stats.at<int>(selectedComponent, cv::CC_STAT_AREA)) {
-        selectedComponent = i;
-      }
-    }
-  }
-  cv::Mat mask = cv::Mat::zeros(img.size(), CV_8UC1);
-  if (selectedComponent != -1) {
-    mask = (labels == selectedComponent) / 255;
-  }
-  cv::Mat resultImage = cv::Mat::zeros(img.size(), img.type());
-  img.copyTo(resultImage, mask);
-  cv::bitwise_not(resultImage, resultImage);
-  return resultImage;
-}
-
-@end

package/ios/RnExecutorch/models/stt/Moonshine.hpp

@@ -1,13 +0,0 @@
-#ifndef Moonshine_hpp
-#define Moonshine_hpp
-
-#import "MoonshineEncoder.hpp"
-#import "Moonshinedecoder.hpp"
-#import "SpeechToTextBaseModel.hpp"
-#import <Foundation/Foundation.h>
-#import <UIKit/UIKit.h>
-
-@interface Moonshine : SpeechToTextBaseModel
-@end
-
-#endif /* Moonshine_hpp */

package/ios/RnExecutorch/models/stt/Moonshine.mm

@@ -1,64 +0,0 @@
-#import "Moonshine.hpp"
-
-@implementation Moonshine {
-  MoonshineEncoder *encoder;
-  MoonshineDecoder *decoder;
-  NSNumber *START_TOKEN;
-  NSNumber *EOS_TOKEN;
-  NSArray *encoderLastHiddenState;
-  int maxSeqLen;
-}
-
-- (instancetype)init {
-  self = [super init];
-  if (self) {
-    START_TOKEN = @1;
-    EOS_TOKEN = @2;
-    maxSeqLen = 180;
-  }
-  return self;
-}
-
-- (NSArray *)encode:(NSArray *)waveform {
-  self->encoderLastHiddenState = [self->encoder encode:waveform];
-
-  if (!self->encoderLastHiddenState) {
-    [NSException raise:@"forward_error" format:nil];
-  }
-  return self->encoderLastHiddenState;
-}
-
-- (NSArray *)decode:(NSArray *)prevTokens
-    encoderLastHiddenState:(NSArray *)encoderLastHiddenState {
-  if ([encoderLastHiddenState count] > 0) {
-    return [self->decoder decode:prevTokens
-          encoderLastHiddenState:encoderLastHiddenState];
-  }
-  return [self->decoder decode:prevTokens
-        encoderLastHiddenState:self->encoderLastHiddenState];
-}
-
-- (void)loadModules:(NSArray *)modelSources {
-
-  self->encoder = [[MoonshineEncoder alloc] init];
-  self->decoder = [[MoonshineDecoder alloc] init];
-
-  // Load encoder
-  [self loadModuleHelper:self->encoder
-      withSource:[modelSources objectAtIndex:0]
-      onSuccess:^{
-        // Load decoder after encoder
-        [self loadModuleHelper:self->decoder
-            withSource:[modelSources objectAtIndex:1]
-            onSuccess:^{
-            }
-            onFailure:^(NSString *errorCode) {
-              [NSException raise:@"init_decoder_error" format:@"%d", errorCode];
-            }];
-      }
-      onFailure:^(NSString *errorCode) {
-        [NSException raise:@"init_encoder_error" format:@"%d", errorCode];
-      }];
-}
-
-@end

package/ios/RnExecutorch/models/stt/MoonshineDecoder.hpp

@@ -1,16 +0,0 @@
-#ifndef MoonshineDecoder_hpp
-#define MoonshineDecoder_hpp
-
-#import "../BaseModel.h"
-#import "ExecutorchLib/ETModel.h"
-#import <Foundation/Foundation.h>
-#import <UIKit/UIKit.h>
-
-@interface MoonshineDecoder : BaseModel
-
-- (NSArray *)decode:(NSArray *)prevTokens
-    encoderLastHiddenState:(NSArray *)encoderLastHiddenState;
-
-@end
-
-#endif /* MoonshineDecoder_hpp */

package/ios/RnExecutorch/models/stt/MoonshineDecoder.mm

@@ -1,24 +0,0 @@
-#import "MoonshineDecoder.hpp"
-#import "../../utils/ScalarType.h"
-
-@implementation MoonshineDecoder
-
-- (NSArray *)decode:(NSArray *)prevTokens
-    encoderLastHiddenState:(NSArray *)encoderLastHiddenState {
-  NSNumber *innerDim = @288;
-  NSNumber *noEncoderTokens =
-      @(@([encoderLastHiddenState count]).intValue / [innerDim intValue]);
-  NSArray *encoderLastHiddenStateShape = @[ @1, noEncoderTokens, innerDim ];
-  NSArray *prevTokensShape = @[ @1, @([prevTokens count]) ];
-  NSArray *predictedToken = [self
-         execute:@"forward_cached"
-          inputs:@[
-            prevTokens, [NSArray arrayWithObjects:encoderLastHiddenState, nil]
-          ]
-          shapes:@[ prevTokensShape, encoderLastHiddenStateShape ]
-      inputTypes:@[ ScalarType.Long, ScalarType.Float ]];
-
-  return [[predictedToken objectAtIndex:0] lastObject];
-}
-
-@end

package/ios/RnExecutorch/models/stt/MoonshineEncoder.hpp

@@ -1,15 +0,0 @@
-#ifndef MoonshineEncoder_hpp
-#define MoonshineEncoder_hpp
-
-#import "../BaseModel.h"
-#import "ExecutorchLib/ETModel.h"
-#import <Foundation/Foundation.h>
-#import <UIKit/UIKit.h>
-
-@interface MoonshineEncoder : BaseModel
-
-- (NSArray *)encode:(NSArray *)waveform;
-
-@end
-
-#endif /* MoonshineEncoder_hpp */

package/ios/RnExecutorch/models/stt/MoonshineEncoder.mm

@@ -1,18 +0,0 @@
-#import "MoonshineEncoder.hpp"
-#import "../../utils/ScalarType.h"
-
-NSArray *waveformTypes = [NSArray arrayWithObject:ScalarType.Float];
-
-@implementation MoonshineEncoder
-
-- (NSArray *)encode:(NSArray *)waveform {
-  NSArray *waveformShape =
-      [NSArray arrayWithObject:@[ @1, @([waveform count]) ]];
-  NSArray *result = [self forward:[NSArray arrayWithObject:waveform]
-                           shapes:waveformShape
-                       inputTypes:waveformTypes];
-
-  return [result objectAtIndex:0];
-}
-
-@end

package/ios/RnExecutorch/models/stt/SpeechToTextBaseModel.hpp

@@ -1,26 +0,0 @@
-#ifndef SpeechToTextBaseModel_hpp
-#define SpeechToTextBaseModel_hpp
-
-#import "../BaseModel.h"
-#import <Foundation/Foundation.h>
-#import <UIKit/UIKit.h>
-@interface SpeechToTextBaseModel : NSObject {
-@public
-  NSNumber *START_TOKEN;
-  NSNumber *EOS_TOKEN;
-  int maxSeqLen;
-  int fftSize;
-}
-
-- (NSArray *)encode:(NSArray *)waveform;
-- (NSArray *)decode:(NSArray *)prevTokens
-    encoderLastHiddenState:(NSArray *)encoderLastHiddenState;
-- (void)loadModules:(NSArray *)modelSources;
-- (void)loadModuleHelper:(id)model
-              withSource:(NSString *)source
-               onSuccess:(void (^)(void))success
-               onFailure:(void (^)(NSString *))failure;
-
-@end
-
-#endif /* SpeechToTextBaseModel_hpp */

package/ios/RnExecutorch/models/stt/SpeechToTextBaseModel.mm

@@ -1,19 +0,0 @@
-#import "SpeechToTextBaseModel.hpp"
-
-@implementation SpeechToTextBaseModel
-
-- (void)loadModuleHelper:(BaseModel *)model
-              withSource:(NSString *)source
-               onSuccess:(void (^)(void))success
-               onFailure:(void (^)(NSString *))failure {
-  NSNumber *errorCode = [model loadModel:source];
-
-  if ([errorCode intValue] != 0) {
-    failure([NSString stringWithFormat:@"%ld", (long)[errorCode longValue]]);
-    return;
-  }
-
-  success();
-}
-
-@end

package/ios/RnExecutorch/models/stt/Whisper.hpp

@@ -1,12 +0,0 @@
-#ifndef Whisper_hpp
-#define Whisper_hpp
-
-#import "ExecutorchLib/ETModel.h"
-#import "SpeechToTextBaseModel.hpp"
-#import <Foundation/Foundation.h>
-#import <UIKit/UIKit.h>
-
-@interface Whisper : SpeechToTextBaseModel
-@end
-
-#endif /* Whisper_hpp */