I’m working on an image classification project for iOS and I’m a bit stuck on picking the right model. I’ve read a bunch of articles, but I’m still not sure. Here’s what I’m looking for:
- Not too fast, super accurate: The model can be slow (like, up to 30 seconds slow), but it needs to be really accurate.
- Size matters: It should be under 300 MB.
- iOS-friendly: Compatibility with iOS’s Neural Engine would be a huge plus.
Some options that I’ve found:
- Inception V3
- Inception V4
- FBNet
- MobileNet V3
- EfficientNet Lite (Lite, because of the Neural Engine support)
I’d really appreciate your thoughts on these models, especially if you’ve used them before. Here’s what I want to know:
Which one’s the most accurate for image classification?
Any trouble running these on iOS?
If you think there’s a better model out there, I’m all ears.Just let me know why you think it’s better.
Thanks a ton!
