Python: I receive an error while running trainer for my facial recognition program in windows 10 but it work on ubuntu 20.04

I would like to use a training for my facial recognition program I receive an error I tried everything but I can't find the solution , in this project even with the same setting on ubuntu 20.04 it worked for me but when I migrated to windows 10, I left the same setting(library) and I got the error.

Another project which uses cnn and hog support works very well with dlib, face-recognition and opencv on ubuntu 20.04 et windows 10 I don't know why in this project it does not work in windows 10.

versions:

os == windows 10
python == 3.6
onnx==1.6.0
onnx-tf==1.3.0
onnxruntime==0.5.0
opencv-python==4.1.1.26
tensorflow==1.13.1
dlib==19.19.0

script (trainer.py):

import cv2
import numpy as np
import onnx
import onnxruntime as ort
from onnx_tf.backend import prepare
import os
import dlib
from imutils import face_utils
import tensorflow as tf
import pickle


def area_of(left_top,right_bottom):
    hw = np.clip(right_bottom-left_top,0.0,None) 
    return hw[...,0]*hw[...,1]


def iou_of(boxes0,boxes1,eps=1e-5):
    overlap_left_top = np.maximum(boxes0[..., :2], boxes1[..., :2])
    overlap_right_bottom = np.minimum(boxes0[..., 2:], boxes1[..., 2:])
    overlap_area = area_of(overlap_left_top, overlap_right_bottom)
    area0 = area_of(boxes0[..., :2], boxes0[..., 2:])
    area1 = area_of(boxes1[..., :2], boxes1[..., 2:])
    return overlap_area / (area0 + area1 - overlap_area + eps)


def hard_nms(box_scores,iou_threshold,top_k=-1,candidate_size=200):
    scores = box_scores[:, -1]
    boxes = box_scores[:, :-1]
    picked = []

    indexes = np.argsort(scores)
    indexes = indexes[-candidate_size:]
    while len(indexes) > 0:
        current = indexes[-1]

        picked.append(current)
        if 0 < top_k == len(picked) or len(indexes) == 1:
            break
        current_box = boxes[current, :]
        indexes = indexes[:-1]
        rest_boxes = boxes[indexes, :]
        ##Return intersection-over-union (Jaccard index) of boxes.
        iou = iou_of(
            rest_boxes,
            np.expand_dims(current_box, axis=0),
        )

        indexes = indexes[iou <= iou_threshold]
    return box_scores[picked, :]


def predict(width, height, confidences, boxes, prob_threshold, iou_threshold=0.5, top_k=-1):
    boxes = boxes[0]
    confidences = confidences[0]
    picked_box_probs = []
    picked_labels = []

    for class_index in range(1, confidences.shape[1]):
        probs = confidences[:, class_index]
        mask = probs > prob_threshold
        probs = probs[mask]
        if probs.shape[0] == 0:
            continue
        subset_boxes = boxes[mask, :]
        #joignez une séquence de tableaux le long d'un axe existant.
        #numpy.reshape(ligne,colonne):Donne une nouvelle forme à un tableau sans modifier ses données.
        box_probs = np.concatenate([subset_boxes, probs.reshape(-1, 1)], axis=1)
        # Perform hard non-maximum-supression to filter out boxes with iou greater than threshold
        box_probs = hard_nms(box_probs,
                             iou_threshold=iou_threshold,
                             top_k=top_k,
                             )
        picked_box_probs.append(box_probs)

        picked_labels.extend([class_index] * box_probs.shape[0])
    if not picked_box_probs:
        return np.array([]), np.array([]), np.array([])
    picked_box_probs = np.concatenate(picked_box_probs)
    picked_box_probs[:, 0] *= width
    picked_box_probs[:, 1] *= height
    picked_box_probs[:, 2] *= width
    picked_box_probs[:, 3] *= height

    return picked_box_probs[:, :4].astype(np.int32), np.array(picked_labels), picked_box_probs[:, 4]

onnx_path = 'ultra_light_640.onnx'
onnx_model = onnx.load(onnx_path)
predictor = prepare(onnx_model)
ort_session = ort.InferenceSession(onnx_path)
input_name = ort_session.get_inputs()[0].name

fa = face_utils.facealigner.FaceAligner(shape_predictor, desiredFaceWidth=112, desiredLeftEye=(0.3, 0.3))


TRAINING_BASE = 'faces/training/'
dirs = os.listdir(TRAINING_BASE)

# images and names for later use
images = []
names = []

for label in dirs:
    for i, fn in enumerate(os.listdir(os.path.join(TRAINING_BASE, label))):
        print(f"start collecting faces from {label}'s data")
        cap = cv2.VideoCapture(os.path.join(TRAINING_BASE, label, fn))
        frame_count = 0
        while True:
            # read video frame
            ret, raw_img = cap.read()

            if frame_count % 5 == 0 and raw_img is not None:
                # size photo
                h, w, _ = raw_img.shape
                # conversion d'espace colorimétrique.
                img = cv2.cvtColor(raw_img, cv2.COLOR_BGR2RGB)
                # Redimmensionner une image
                img = cv2.resize(img, (640, 480))
                img_mean = np.array([127, 127, 127])
                img = (img - img_mean) / 128
                img = np.transpose(img, [2, 0, 1])
                img = np.expand_dims(img, axis=0)
                img = img.astype(np.float32)

                confidences, boxes = ort_session.run(None, {input_name: img})

                boxes, labels, probs = predict(w, h, confidences, boxes, 0.7)

                if boxes.shape[0] > 0:
                    x1, y1, x2, y2 = boxes[0, :]

                    gray = cv2.cvtColor(raw_img, cv2.COLOR_BGR2GRAY)

                    aligned_face = fa.align(raw_img, gray, dlib.rectangle(left=int(x1), top=int(y1), right=int(x2), bottom=int(y2)))
                    aligned_face = cv2.resize(aligned_face, (112, 112))

                    cv2.imwrite(f'faces/tmp/{label}_{frame_count}.jpg', aligned_face)

                    aligned_face = aligned_face - 127.5
                    aligned_face = aligned_face * 0.0078125
                    images.append(aligned_face)
                    names.append(label)

            frame_count += 1
            # if video end
            if frame_count == cap.get(cv2.CAP_PROP_FRAME_COUNT):
                break


with tf.Graph().as_default():
    with tf.Session() as sess:
        print("loading checkpoint ...")
        saver = tf.train.import_meta_graph('models/mfn/m1/mfn.ckpt.meta')
        saver.restore(sess, 'models/mfn/m1/mfn.ckpt')

        images_placeholder = tf.get_default_graph().get_tensor_by_name("input:0")
        embeddings = tf.get_default_graph().get_tensor_by_name("embeddings:0")
        phase_train_placeholder = tf.get_default_graph().get_tensor_by_name("phase_train:0")
        embedding_size = embeddings.get_shape()[1]

        feed_dict = {images_placeholder: images, phase_train_placeholder: False}
        embeds = sess.run(embeddings, feed_dict=feed_dict)
        print("saving embeddings")
        with open("embeddings/embeddings.pkl", "wb") as f:
            pickle.dump((embeds, names), f)
        print("Done !")'''

Error message:

G:\PROJET\PYTHON\FACE_RECOGNITION\Real_time_face_recognition_with_CPU\venv\Scripts\python.exe G:/PROJET/PYTHON/FACE_RECOGNITION/Real_time_face_recognition_with_CPU/trainer.py
G:\PROJET\PYTHON\FACE_RECOGNITION\Real_time_face_recognition_with_CPU\venv\lib\site-packages\tensorflow\python\framework\dtypes.py:526: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
  _np_qint8 = np.dtype([("qint8", np.int8, 1)])
G:\PROJET\PYTHON\FACE_RECOGNITION\Real_time_face_recognition_with_CPU\venv\lib\site-packages\tensorflow\python\framework\dtypes.py:527: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
  _np_quint8 = np.dtype([("quint8", np.uint8, 1)])
G:\PROJET\PYTHON\FACE_RECOGNITION\Real_time_face_recognition_with_CPU\venv\lib\site-packages\tensorflow\python\framework\dtypes.py:528: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
  _np_qint16 = np.dtype([("qint16", np.int16, 1)])
G:\PROJET\PYTHON\FACE_RECOGNITION\Real_time_face_recognition_with_CPU\venv\lib\site-packages\tensorflow\python\framework\dtypes.py:529: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
  _np_quint16 = np.dtype([("quint16", np.uint16, 1)])
G:\PROJET\PYTHON\FACE_RECOGNITION\Real_time_face_recognition_with_CPU\venv\lib\site-packages\tensorflow\python\framework\dtypes.py:530: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
  _np_qint32 = np.dtype([("qint32", np.int32, 1)])
G:\PROJET\PYTHON\FACE_RECOGNITION\Real_time_face_recognition_with_CPU\venv\lib\site-packages\tensorflow\python\framework\dtypes.py:535: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
  np_resource = np.dtype([("resource", np.ubyte, 1)])
G:\PROJET\PYTHON\FACE_RECOGNITION\Real_time_face_recognition_with_CPU\venv\lib\site-packages\onnx_tf\common\__init__.py:87: UserWarning: onnx_tf.common.get_outputs_names is deprecated. It will be removed in future release. Use TensorflowGraph.get_outputs_names instead.
  warnings.warn(message)

WARNING: The TensorFlow contrib module will not be included in TensorFlow 2.0.
For more information, please see:
  * https://github.com/tensorflow/community/blob/master/rfcs/20180907-contrib-sunset.md
  * https://github.com/tensorflow/addons
If you depend on functionality not listed there, please file an issue.

G:\PROJET\PYTHON\FACE_RECOGNITION\Real_time_face_recognition_with_CPU\venv\lib\site-packages\onnx_tf\common\handler_helper.py:37: UserWarning: Unknown op ConstantFill in domain `ai.onnx`.
  handler.ONNX_OP, handler.DOMAIN or "ai.onnx"))
G:\PROJET\PYTHON\FACE_RECOGNITION\Real_time_face_recognition_with_CPU\venv\lib\site-packages\onnx_tf\common\handler_helper.py:37: UserWarning: Unknown op ImageScaler in domain `ai.onnx`.
  handler.ONNX_OP, handler.DOMAIN or "ai.onnx"))

Process finished with exit code -1073741819 (0xC0000005)