3 ms·
You're probably going to get much better performance if you convert your TensorFlow model to CoreML (use tfcoreml). Also has the advantage of not needing to com
by gok 8y ago
You're probably going to get much better performance if you convert your TensorFlow model to CoreML (use tfcoreml). Also has the advantage of not needing to compile TF for your phone yourself, and distribute it with your app.
- Q6T46nT668w6i3m 8y agoIf, and only if, it can be compiled to CoreML. (I wrote and maintain Keras-RCNN. It does not compile to CoreML. Yet.)
- minimaxir 8y agoOut of curiosity, why does compiling to CoreML fail?
- asparagui 8y agoNot every op supported in generic Tensorflow (software) can be converted to CoreML ops (hardware), which means a number of the more complicated models can't be automagically converted. If you have specific domain knowledge you can add substitute ops/roll your own, but this is beyond the scope of most people trying to bring models over to mobile.
- natch 8y agoMany models exist in many formats that can be converted to CoreML. So that shouldn’t be a problem. Also there is always the option of training a new model from scratch or with transfer learning.
- darylrowland 8y agoWe wrote a similar article to this about a year ago, using CoreML & TuriCreate to train the data... https://medium.com/jigsaw-xyz/realtime-image-recognition-in-reactnative-with-coreml-24ed7d414f23 https://medium.com/jigsaw-xyz/realtime-image-recognition-in-...
- prashp 8y agoTuriCreate also provides good tutorials on doing this: https://apple.github.io/turicreate/docs/userguide/object_detection/ https://apple.github.io/turicreate/docs/userguide/object_det...
- joeblau 8y agoI came here to say the same thing :). It's probably also easier to train using CreateML, but if you already know how to use TensorFlow then you don't really need to use CreateML.
- pj_mukh 8y agoI've heard through the grapevine that Tensorflow-Lite just makes a CoreML conversion for you? Is that true?