$ pip install tf2onnx
$ python -m tf2onnx.convert --tflite face_detection_short_range.tflite --output face_detection_short_range.onnx --opset 13

import torch
import torch.nn as  nn
import numpy as np

emo_affectnet = ResNet50(7, channels=3)
emo_affectnet.load_state_dict(torch.load('FER_static_ResNet50_AffectNet.pt'))
emo_affectnet.eval()

model_cpu = emo_affectnet.cpu()
# cur_face would be cropped face image type of numpy array.
model_traced = torch.jit.trace(model_cpu, (cur_face))

np_cur_face = cur_face.detach().numpy()
np.save("data/cur_face.npy", np_cur_face)

output_model_path = f"models/FER_static_ResNet50_AffectNet_traced.pt"
torch.jit.save(model_traced, output_model_path)

# (1) Get mlange_gen
$ wget <https://github.com/zetic-ai/ZETIC_MLange_document/raw/main/bin/mlange_gen> && chmod 755mlange_gen

# (2) Run mlange_gen for two models
#    - Face detection model
$ ./mlange_gen -m face_detection_short_range.onnx -i input.npy

#    - Face emotion recognition model
$ ./mlange_gen -m FER_static_ResNet50_AffectNet.pt -i

...
MLange Model Key : {YOUR_FACE_DETECTION_MODEL_KEY}
...

...
MLange Model Key : {YOUR_FACE_EMOTION_RECOGNITION_MODEL_KEY}
...

val faceEmotionRecognitionModel = ZeticMLangeModel(this, 'face_emotion_recognition')

faceEmotionRecognitionModel.run(inputs)

val outputs = faceEmotionRecognitionModel.outputBuffers

let faceEmotionRecognitionModel = ZeticMLangeModel('face_emotion_recognition')

faceEmotionRecognitionModel.run(inputs)

let outputs = faceEmotionRecognitionModel.getOutputDataArray()

// (0) Initialize ZeticMLangeFeatureFaceEmotionRecognition
val feature = ZeticMLangeFeatureFaceEmotionRecognition()

// (1) Preprocess bitmap and get processed float array
val inputs = feature.preprocess(bitmap)

...

// (2) Postprocess to bitmap
val resultBitmap = feature.postprocess(outputs

import ZeticMLange

// (0) Initialize ZeticMLangeFeatureFaceEmotionRecognition
let feature = ZeticMLangeFeatureFaceEmotionRecognition()

// (1) Preprocess UIImage and get processed float array
let inputs = feature.preprocess(image)

...

// (2) Postprocess to UIImage
let resultBitmap = feature.postprocess(&outputs)

// (0) Initialization Models
val faceDetectionModel = ZeticMLangeModel(this, 'face_detection')

// (1) Initialization Feature
val faceDetectionFeature = ZeticMLangeFeatureFaceDetection()

// (2) Preprocess Image
val faceDetectionInputs = faceDetectionFeature.preprocess(bitmap)

// (3) Process Model
faceDetectionModel.run(faceDetectionInputs)
val faceDetectionOutputs = faceDetectionModel.outputBuffers

// (4) Postprocess model run result
val faceDetectionPostprocessed = faceDetectionFeature.postprocess(faceDetectionOutputs

// (0) Initialization Models
val faceEmotionRecognitionModel = ZeticMLangeModel(this, 'face_emotion_recognition')

// (1) Initialization Feature
val faceEmotionRecognitionFeature = ZeticMLangeFeatureFaceEmotionRecognition()

// (2) Preprocess Image
val faceEmotionRecognitionInputs = faceEmotionRecognitionFeature.preprocess(bitmap, faceDetectionPostprocessed)

// (3) Process Model
faceEmotionRecognitionModel.run(faceEmotionRecognitionInputs)
val faceEmotionRecognitionOutputs = faceEmotionRecognitionModel.outputBuffers

// (4) Postprocess model run result
val faceEmotionRecognitionPostprocessed = faceEmotionRecognitionFeature.postprocess(faceEmotionRecognitionOutputs

// (0) Initialization Models
let faceEmotionRecognitionModel = ZeticMLangeModel('face_emotion_recognition')

// (1) Initialization Feature
let faceEmotionRecognitionFeature = ZeticMLangeFeatureFaceEmotionRecognition()

// (2) Preprocess Image
let faceEmotionRecognitionInputs = faceEmotionRecognitionFeature.preprocess(bitmap, faceDetectionPostprocessed)

// (3) Process Model
faceEmotionRecognitionModel.run(faceEmotionRecognitionInputs)
let faceEmotionRecognitionOutputs = faceEmotionRecognitionModel.outputBuffers

// (4) Postprocess model run result
let faceEmotionRecognitionPostprocessed = faceEmotionRecognitionFeature.postprocess(&faceEmotionRecognitionOutputs)

// (0) Initialization Models
let faceDetectionModel = ZeticMLangeModel('face_detection')

// (1) Initialization Feature
let faceDetectionFeature = ZeticMLangeFeatureFaceDetection()

// (2) Preprocess Image
let faceDetectionInputs = faceDetectionFeature.preprocess(bitmap)

// (3) Process Model
faceDetectionModel.run(faceDetectionInputs)
let faceDetectionOutputs = faceDetectionModel.outputBuffers

// (4) Postprocess model run result
let faceDetectionPostprocessed = faceDetectionFeature.postprocess(&faceDetectionOutputs)