Core ML model 资源

1、苹果官方提供的model

modelMobileNet

MobileNets are based on a streamlined architecture that have depth-wise separable convolutions to build lightweight, deep neural networks. Detects the dominant objects present in an image from a set of 1000 categories such as trees, animals, food, vehicles, people, and more.
View original model details
Download Core ML Model

SqueezeNet

Detects the dominant objects present in an image from a set of 1000 categories such as trees, animals, food, vehicles, people, and more. With an overall footprint of only 5 MB, SqueezeNet has a similar level of accuracy as AlexNet but with 50 times fewer parameters.
View original model details
Download Core ML Model

Places205-GoogLeNet

Detects the scene of an image from 205 categories such as an airport terminal, bedroom, forest, coast, and more.
View original model details
Download Core ML Model

ResNet50

Detects the dominant objects present in an image from a set of 1000 categories such as trees, animals, food, vehicles, people, and more.
View original model details
Download Core ML Model

Inception v3

Detects the dominant objects present in an image from a set of 1000 categories such as trees, animals, food, vehicles, people, and more.
View original model details
Download Core ML Model

VGG16

Detects the dominant objects present in an image from a set of 1000 categories such as trees, animals, food, vehicles, people, and more.
View original model details
Download Core ML Model

2、第三方共享的model

Awesome-CoreML-Models

2.1 Image Processing

Models that takes image data as input and output useful information about the image.

MobileNet

The network from the paper 'MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications', trained on the ImageNet dataset.
Download | Demo | Reference

GoogLeNetPlaces

Detects the scene of an image from 205 categories such as airport, bedroom, forest, coast etc.
Download | Demo | Reference

Inceptionv3

Detects the dominant objects present in an image from a set of 1000 categories such as trees, animals, food, vehicles, person etc. The top-5 error from the original publication is 5.6%.
Download | Demo | Reference

Resnet50

Detects the dominant objects present in an image from a set of 1000 categories such as trees, animals, food, vehicles, person etc. The top-5 error from the original publication is 7.8%.
Download | Demo | Reference

VGG16

Detects the dominant objects present in an image from a set of 1000 categories such as trees, animals, food, vehicles, person etc. The top-5 error from the original publication is 7.4%.
Download | Demo | Reference

CarRecognition

Predict the brand & model of a car.
Download | Demo | Reference

TinyYOLO

The Tiny YOLO network from the paper 'YOLO9000: Better, Faster, Stronger' (2016), arXiv:1612.08242
Download | Demo | Reference

AgeNet

Age Classification using Convolutional Neural Networks
Download | Demo | Reference

GenderNet

Gender Classification using Convolutional Neural Networks
Download | Demo | Reference

MNIST

Predicts a handwritten digit.
Download | Demo | Reference

CNNEmotions

Emotion Recognition in the Wild via Convolutional Neural Networks and Mapped Binary Patterns
Download | Demo | Reference

VisualSentimentCNN

Fine-tuning CNNs for Visual Sentiment Prediction
Download | Demo | Reference

Food101

This model takes a picture of a food and predicts its name
Download | Demo | Reference

Oxford102

Classifying images in the Oxford 102 flower dataset with CNNs
Download | Demo | Reference

FlickrStyle

Finetuning CaffeNet on Flickr Style
Download | Demo | Reference

RN1015k500

Predict the location where a picture was taken.
Download | Demo | Reference

Nudity

Classifies an image either as NSFW (nude) or SFW (not nude)
Download | Demo | Reference

2.2、Style Transfer

Models that transform image to specific style.

HED_so

Holistically-Nested Edge Detection. Side outputs
Download | Demo | Reference

FNS-Candy

Feedforward style transfer https://github.com/jcjohnson/fast-neural-style
Download | Demo | Reference

FNS-Feathers

Feedforward style transfer https://github.com/jcjohnson/fast-neural-style
Download | Demo | Reference

FNS-La-Muse

Feedforward style transfer https://github.com/jcjohnson/fast-neural-style
Download | Demo | Reference

FNS-The-Scream

Feedforward style transfer https://github.com/jcjohnson/fast-neural-style
Download | Demo | Reference

FNS-Udnie

Feedforward style transfer https://github.com/jcjohnson/fast-neural-style
Download | Demo | Reference

FNS-Mosaic

Feedforward style transfer https://github.com/jcjohnson/fast-neural-style
Download | Demo | Reference

AnimeScale2x

Process a bicubic-scaled anime-style artwork
Download | Demo | Reference

2.3、Text Analysis

Models that takes text data as input and output useful information about the text.

SentimentPolarity

Sentiment polarity LinearSVC.
Download | Demo | Reference

DocumentClassification

Classify news articles into 1 of 5 categories.
Download | Demo | Reference

MessageClassifier

Detect whether a message is spam.
Download | Demo | Reference

NamesDT

Gender Classification using DecisionTreeClassifier
Download | Demo | Reference

2.4、Others

Exermote

Predicts the exercise, when iPhone is worn on right upper arm.
Download | Demo | Reference

GestureAI

GestureAI
Download | Demo | Reference

引用

Awesome-CoreML-Models