Update README.md

Author: instant-high

README.md

@@ -1,6 +1,10 @@
 # face-parsing-onnx
 face parsing / face masking converted to onnx model
 
+CPU or GPU
+
+Requirements: opencv-python, numpy, onnxruntime (onnruntime-gpu, cudatoolkit=11.2 cudnn=8.1.0)
+
 python inference.py --source_image "image.jpg" --parser_index 1,2,3,4,5,6,10,11,12,13