llsourcell / convolutional_neural_network Goto Github PK

I get the following issue while trying to run using python 3.
ImportError: bad magic number in 'config': b'\x03\xf3\r\n'

What could possibly be causing this issue? Is it because my python 3 version is 3.5.2 instead of 3.5.1 as specified in the runtime.txt file? My default python interpreter is set to python 2 which is the default on a Ubuntu system.

Thanks!

ImportError Traceback (most recent call last)
in ()
1 import pickle #saving and loading our serialized model
2 import numpy as np #matrix math
----> 3 from app.model.preprocessor import Preprocessor as img_prep #image preprocessing
4
5 #class for loading our saved model and classifying new images

C:\Users\DELL\Convolutional_neural_network-master\app_init_.py in ()
6 app.config.from_object('config')
7
----> 8 from app import views

C:\Users\DELL\Convolutional_neural_network-master\app\views.py in ()
1 from flask import render_template, flash, redirect, request, jsonify
2 from app import app
----> 3 from app.model.preprocessor import Preprocessor as img_prep
4 from app.model.alpha_cnn_predict import LiteOCR
5 import json

ImportError: bad magic number in 'app.model': b'\x03\xf3\r\n'

Confused about layer sequence

In the image there is "conv+relu" -> "pooling" -> "conv+relu" -> "pooling"
In the steps there is "conv" -> "pooling" -> "relu" -> "conv" -> "pooling" -> "relu"
In the code there is "conv" -> "relu" -> "conv" -> "relu" > "pooling"

So which one is correct ?

The model is not yet identifying the small alphabets.