amitshekhariitbhu / androidtensorflowmnistexample

Android TensorFlow MachineLearning MNIST Example (Building Model with TensorFlow for Android)

About me

Hi, I am Amit Shekhar, I have taught and mentored many developers, and their efforts landed them high-paying tech jobs, helped many tech companies in solving their unique problems, and created many open-source libraries being used by top companies. I am passionate about sharing knowledge through open-source, blogs, and videos.

Learn from my blogs: amitshekhar.me/blog

Learn from my videos: youtube.com/@amitshekhar

Get High Paying Tech Job: amitshekhar.me

You can connect with me on:

• Twitter
• LinkedIn

androidtensorflowmnistexample's Issues

running Deep Learning Models on Android VS raspberry PI 3

Hello ,
What Are Your Views on running Deep Learning Models on Android VS raspberry PI 3..

I have a Camera Module Attached To raspberry pi 3 which Clicks Images ..Should I use deep Learning model at the Raspberry Pi level or should I send the Image to android and then use the deep learning Model...

The Ultimate result of the OutPUT label shall be Visible on the Android app...

Looking forward to hear from you...

java.io.IOException:Not a valid TensorFlow Graph serialization: NodeDef mentions attr 'dilations'

Repro steps:

1. replace mnist_model_graph.pb in app/src/main/assets with AndroidTensorFlowMNISTExample/model/
2. Run "Debug app" in Android Studio

Observed:
App crash with below log.
E/AndroidRuntime:
FATAL EXCEPTION: pool-1-thread-1 Process: com.mindorks.tensorflowexample, PID: 3776
java.lang.RuntimeException: Error initializing TensorFlow!
at com.mindorks.tensorflowexample.MainActivity$3.run(MainActivity.java:114)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1162)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:636)
at java.lang.Thread.run(Thread.java:764)
Caused by: java.lang.RuntimeException: Failed to load model from 'file:///android_asset/mnist_model_graph.pb'
at org.tensorflow.contrib.android.TensorFlowInferenceInterface.(TensorFlowInferenceInterface.java:100)
at com.mindorks.tensorflowexample.TensorFlowImageClassifier.create(TensorFlowImageClassifier.java:101)
at com.mindorks.tensorflowexample.MainActivity$3.run(MainActivity.java:104)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1162) 
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:636) 
at java.lang.Thread.run(Thread.java:764) 
Caused by: java.io.IOException: Not a valid TensorFlow Graph serialization: NodeDef mentions attr 'dilations' not in Op<name=Conv2D; signature=input:T, filter:T -> output:T; attr=T:type,allowed=[DT_HALF, DT_FLOAT]; attr=strides:list(int); attr=use_cudnn_on_gpu:bool,default=true; attr=padding:string,allowed=["SAME", "VALID"]; attr=data_format:string,default="NHWC",allowed=["NHWC", "NCHW"]>; NodeDef: Conv2D = Conv2D[T=DT_FLOAT, data_format="NHWC", dilations=[1, 1, 1, 1], padding="SAME", strides=[1, 1, 1, 1], use_cudnn_on_gpu=true](Reshape, WC1). (Check whether your GraphDef-interpreting binary is up to date with your GraphDef-generating binary.).

Finding:

1. These 2 mnist_model_graph.pb difference in file size.
2. Is it possible that these 2 .pb file is generated by different tensorflow version causes this error?

Can we train custom model in the Android application, not in python?

Can we train custom model in the Android application, not in python?

When I run the mnist.py, it will cause the computer to crash. Is there any way to solve this problem?