fix: input shape

Author: dev-diaries41

ml/src/main/java/com/fpf/smartscansdk/ml/models/providers/embeddings/clip/ClipImageEmbedder.kt

@@ -45,7 +45,7 @@ class ClipImageEmbedder(private val context: Context, modelSource: ModelSource,
     override suspend fun embed(data: Bitmap): FloatArray = withContext(Dispatchers.Default) {
         if (!isInitialized()) throw IllegalStateException("Model not initialized")
 
-        val inputShape = longArrayOf(DIM_BATCH_SIZE.toLong(), DIM_PIXEL_SIZE.toLong(), IMAGE_SIZE_X.toLong(), IMAGE_SIZE_Y.toLong())
+        val inputShape = longArrayOf(DIM_BATCH_SIZE.toLong(), DIM_PIXEL_SIZE.toLong(), IMAGE_SIZE_Y.toLong(), IMAGE_SIZE_X.toLong())
         val imgData: FloatBuffer = preProcess(data)
         val inputName = model.getInputNames()?.firstOrNull() ?: throw IllegalStateException("Model inputs not available")
         val output = model.run(mapOf(inputName to TensorData.FloatBufferTensor(imgData, inputShape)))

ml/src/main/java/com/fpf/smartscansdk/ml/models/providers/embeddings/dino/DinoV2SmallImageEmbedder.kt

@@ -51,7 +51,7 @@ class DinoV2SmallImageEmbedder(
     override suspend fun embed(data: Bitmap): FloatArray = withContext(Dispatchers.Default) {
         if (!isInitialized()) throw IllegalStateException("Model not initialized")
 
-        val inputShape = longArrayOf(DIM_BATCH_SIZE.toLong(), DIM_PIXEL_SIZE.toLong(), IMAGE_SIZE_X.toLong(), IMAGE_SIZE_Y.toLong())
+        val inputShape = longArrayOf(DIM_BATCH_SIZE.toLong(), DIM_PIXEL_SIZE.toLong(), IMAGE_SIZE_Y.toLong(), IMAGE_SIZE_X.toLong())
         val imgData: FloatBuffer = preProcess(data)
         val inputName = model.getInputNames()?.firstOrNull() ?: throw IllegalStateException("Model inputs not available")
         val output = model.run(mapOf(inputName to TensorData.FloatBufferTensor(imgData, inputShape)))

ml/src/main/java/com/fpf/smartscansdk/ml/models/providers/embeddings/inception/InceptionResnetFaceEmbedder.kt

@@ -46,7 +46,7 @@ class InceptionResnetFaceEmbedder(
         if (!isInitialized()) throw IllegalStateException("Model not initialized")
 
         val imgData = preProcess(data)
-        val inputShape = longArrayOf(DIM_BATCH_SIZE.toLong(), DIM_PIXEL_SIZE.toLong(), IMAGE_SIZE_X.toLong(), IMAGE_SIZE_Y.toLong())
+        val inputShape = longArrayOf(DIM_BATCH_SIZE.toLong(), DIM_PIXEL_SIZE.toLong(), IMAGE_SIZE_Y.toLong(), IMAGE_SIZE_X.toLong())
         val inputName = model.getInputNames()?.firstOrNull() ?: throw IllegalStateException("Model inputs not available")
         val output = model.run(mapOf(inputName to TensorData.FloatBufferTensor(imgData, inputShape)))
         (output.values.first() as Array<FloatArray>)[0]