diff --git a/test_script/test_to_onnx.py b/test_script/test_to_onnx.py
new file mode 100644
index 0000000..d8108b3
--- /dev/null
+++ b/test_script/test_to_onnx.py
@@ -0,0 +1,33 @@
+from os import getcwd
+from os.path import join
+from onnxruntime import InferenceSession, SessionOptions, __version__, get_all_providers, get_available_providers
+from PIL import Image
+import numpy
+
+print(get_all_providers())
+print(get_available_providers())
+
+onnx_session = InferenceSession(
+    path_or_bytes=r"C:\Users\tomokazu\build\retinaface\retinaface_sim.onnx",
+    providers=[
+        'CUDAExecutionProvider',
+        ('TensorrtExecutionProvider', {
+            'trt_engine_cache_enable': True,
+            'trt_engine_cache_path': join(getcwd(), 'onnx_cache'),
+            'trt_fp16_enable': True,
+        }),
+        'DmlExecutionProvider',
+        'CPUExecutionProvider'
+    ]
+)
+print(__version__)
+image_arr = numpy.expand_dims(numpy.array(
+    Image.open(r'C:\Users\tomokazu\CLionProjects\ameba_blog_downloader\manaka_test.jpg').convert('RGB'))
+                              .transpose(2, 0, 1).astype(numpy.float32), 0)
+# image_arr /= 255.0
+print(image_arr)
+print(image_arr.shape)
+res = onnx_session.run(input_feed={'input': image_arr}, output_names=["bbox", "confidence", "landmark"])
+for val in res:
+    print(val)
+    print(val.shape)