Open macinspak opened 2 years ago
To be clear, we are intending to try and match people identified in a photo on one platform against people in different photos on the other platform, so we need to use the same models in both environments.
Thanks again!
Hi, great work and thanks, this library has been an awesome easy inclusion to our solution.
however, now we need to implement the same tensor models to find faces and the 68 landmarks on iOS devices using swift. To do so, we would prefer to convert the appropriate models to coreml using apples python converter.
I’ve been trying to use the true converter to convert the js models back to the original models to load into python, but having a few issues. Could you provide any pointers? Do you have the original tf saved models or know where they are so we aren’t converting backwards?
thanks so much again and keep up the great work!