hollance / tensorflow-ios-example Goto Github PK

View Code? Open in Web Editor NEW

442.0 20.0 86.0 425 KB

Source code for my blog post "Getting started with TensorFlow on iOS"

Shell 0.26% Python 40.53% Objective-C 3.46% Objective-C++ 14.79% Swift 40.97%

tensorflow metal ios machine-learning logistic-regression voice-recognition

tensorflow-ios-example's Introduction

TensorFlow on iOS demo

This is the code that accompanies my blog post Getting started with TensorFlow on iOS.

It uses TensorFlow to train a basic binary classifier on the Gender Recognition by Voice and Speech Analysis dataset.

This project includes the following:

The dataset in the file voice.csv.
Python scripts to train the model with TensorFlow on your Mac.
An iOS app that uses the TensorFlow C++ API to do inference.
An iOS app that uses Metal to do inference using the trained model.

Training the model

To train the model, do the following:

Make sure these are installed: python3, numpy, pandas, scikit-learn, tensorflow.
Run the split_data.py script to divide the dataset into a training set and a test set. This creates 4 new files: X_train.npy, y_train.npy, X_test.npy, and y_test.npy.
Run the train.py script. This trains the logistic classifier and saves the model to /tmp/voice every 10,000 training steps. Training happens in an infinite loop and goes on forever, so press Ctrl+C when you're happy with the training set accuracy and the loss no longer becomes any lower.
Run the test.py script to compute the accuracy on the test set. This also prints out a report with precision / recall / f1-score and a confusion matrix.

Using the model with the iOS TensorFlow app

To run the model on the iOS TensorFlow app, do the following:

Clone TensorFlow and build the iOS library.
Open the VoiceTensorFlow Xcode project. In Build Settings, Other Linker Flags and Header Search Paths, change the paths to your local installation of TensorFlow.

The model is already included in the app as inference.pb. If you train the model with different settings, you need to run the freeze_graph and optimize_for_inference tools to create a new inference.pb.

Using the model with the iOS Metal app

To run the model on the iOS Metal app, do the following:

Run the export_weights.py script. This creates two new files that contain the model's learned parameters: W.bin for the weights and b.bin for the bias.
Copy W.bin and b.bin into the VoiceMetal Xcode project and build the app.

You need to run the Metal app on a device, it won't work in the simulator.

tensorflow-ios-example's People

Contributors

Stargazers

Watchers

Forkers

soledad89 terriermon clcarwin chagge benjamesbabala supercomputra achigeor johndpope vskabelkin panthole lcklozz13 wanjinchang tjumick ioshome snakejenny windfringe xiaobailong940 ubtechrobotics shengyang998 shuaiwang007 inureyes ofirbb ruyiwei-cas dimroc svpwm xyxjn abrbrazj zgsxwsdxg rosssong ferasos lnison biranchi2018 schettino isaac234517 ahuang1900 alldev0825 yeison shiuh-yaw cattlelovegrass jbloit hsd315 reversescale prosperityai hctoolsprojects yeshuanova misanyrock kwccoin casillas-qf hhuahh detailscool vistarsoft candidey jiaoshengha 91haba ddviper shaoweistar eaglep91 nsblacker sanguin amandeep-ios embwireless huangpu1 vvvictorlee shenhzou654321 jinyouli boljonggo gail9527 ideal-yang maxprog jordanekm xuyushiguang wolfalonelj vendetta9527 wangdeyu noid2015 colinsongf wzyuzjb rajmanikush debugspy sky9743 linkonbsmrstu kidzss armangorjipoor

tensorflow-ios-example's Issues

Running error

Hi i am getting this error Error: This device does not support Metal Performance Shaders: file /Users/mac/Downloads/TensorFlow-iOS-Example-master 3/VoiceMetal/VoiceMetal/ViewController.swift, line 49

'tensorflow/core/public/session.h' file not found

Hi.

I have run tensorflow/contrib/makefile/build_all_ios.sh and the following files now exist libtensorflow-core.a, libprotobuf.a, libprotobuf-lite.a at:

"/Users/anton6/tensorflow/tensorflow/contrib/makefile/gen/protobuf_ios/lib/libprotobuf-lite.a"

"/Users/anton6/tensorflow/tensorflow/contrib/makefile/gen/protobuf_ios/lib/libprotobuf.a"

"/Users/anton6/tensorflow/tensorflow/contrib/makefile/gen/lib/libtensorflow-core.a"

I have updated the paths in Build Settings/Other Linker Flags accordingly, but I get the error 'tensorflow/core/public/session.h' file not found when trying to run.

Any idea what I am doing wrong?

Thanks

you would need to add code that records audio and then extracts the acoustic properties from the audio.

HI~
Great blog here.
http://machinethink.net/blog/tensorflow-on-ios/?utm_source=Swift_Developments&utm_medium=email&utm_campaign=Swift_Developments_Issue_79

And how to get a mp3 file's properties like your maleExample data ? as your said 'you would first need to convert the audio into these 20 acoustic features.'

hello!it's me again(awkward)!

In your article, you said
Then to install the package:

pip3 uninstall tensorflow
sudo -H pip3 install /tmp/tensorflow_pkg/tensorflow-1.0.0-XXXXXX.whl

I wanna know why I have to uninstall my tensor flow ,if I uninstall it, how can I use it after that?

hello!it's me again, thanks for the guide u gave me before and I fixed that bug,but I met another problem!

crissallandeMacBook-Pro:~ crissallan$ python3 train.py
Training set size: (2217, 20)
W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use SSE4.1 instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use SSE4.2 instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use AVX instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use AVX2 instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use FMA instructions, but these are available on your machine and could speed up CPU computations.
Initial loss 0.693146

hello!I meet a bug when run train.py could you please tell me how to fix it?

excuse me, how to configure paths

hi, i wanna run this project on my iPhone but meet some problems

None of your accounts are a member of 'MGCEKGP4Y4': An unexpected error occurred. Xcode cannot find a team matching 'MGCEKGP4Y4'.

No profiles for 'com.hollance.VoiceTensorFlow' were found: Xcode couldn't find a provisioning profile matching 'com.hollance.VoiceTensorFlow'.

Code signing is required for product type 'Application' in SDK 'iOS 10.3'
problem No.2

Integrating this in swift giving 'string' file not found error

I am using this framework in a swift project and used a bridging header to implement some functionalities. I was included this framework with CocoaPods into my swift project. But while trying to build the project, there are so many syntax errors like #include not found, #include not found etc. I have been banging my head around this issue from past two days but no luck. The Xcode compiler is able to build if i am using an Objective C Project. Somebody please guide me to resolve the issue with swift project

Thanks in Advance

#import <tensorflow/core/public/session.h> error