Tensorflow MNIST and preprocessing
This project have been forked from the excellent project tensorflow-mnist by Ole Kröger: https://medium.com/@o.kroeger/tensorflow-mnist-and-your-own-handwritten-digits-4d1cd32bbab4 https://github.com/opensourcesblog/tensorflow-mnist I appreciate all the work he has done for it and I want to acknowledge it.
This is a project familiarizing myself with tensorflow and the MNIST database.
For me to be able to run this software, there were some steps I had to take: (This is from my installation notes) Install 64-bit Python 3.5 (minimum): Python35 64-bit https://www.python.org/ftp/python/3.5.4/python-3.5.4-amd64-webinstall.exe install for all users Python35 to PATH Install debug symbols
Install tensorflow, pandas and numpy: In a command window (do it anywhere after Python35 is in PATH): pip3 install --update tensorflow pip install pandas pip install numpy
Convert GIT-files from Python2 into Python3: CommandPrompt: > cd C:\GIT\github\tensorflow-mnist.git\ > python C:\Program Files\Python35\Tools\scripts\2to3.py -w mnist.py > python C:\Program Files\Python35\Tools\scripts\2to3.py -w step2.py > python C:\Program Files\Python35\Tools\scripts\2to3.py -w input_data.py > python C:\Program Files\Python35\Tools\scripts\2to3.py -w learn_extra.pyInstall OpenCV: https://solarianprogrammer.com/2016/09/17/install-opencv-3-with-python-3-on-windows/ https://support.microsoft.com/en-us/help/2977003/the-latest-supported-visual-c-downloads vc_redist.x64.exe https://www.lfd.uci.edu/~gohlke/pythonlibs/#opencv opencv_python-3.4.3-cp35-cp35m-win_amd64.whl pip install "opencv_python-3.4.3-cp35-cp35m-win_amd64.whl"
There was an error: "TypeError: only integer scalar arrays can be converted to a scalar index" LINE33 bytestream.read https://stackoverflow.com/questions/42128830/typeerror-only-integer-scalar-arrays-can-be-converted-to-a-scalar-index Change: From: def _read32(bytestream): dt = numpy.dtype(numpy.uint32).newbyteorder('>') return numpy.frombuffer(bytestream.read(4), dtype=dt) To: def _read32(bytestream): dt = numpy.dtype(numpy.uint32).newbyteorder('>') return numpy.frombuffer(bytestream.read(4), dtype=dt)[0]
Don't mind the warning: 2018-09-19 09:45:54.456223: I T:\src\github\tensorflow\tensorflow\core\platform\cpu_feature_guard.cc:141] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 https://stackoverflow.com/questions/47068709/your-cpu-supports-instructions-that-this-tensorflow-binary-was-not-compiled-to-u
mnist.py:
SUCCESS will be in the form of something like this:
0.9145
[8 0 4 3]
1.0
predict_interface_usage.py: A photograph of handwritten digits as input (/img/), SUCCESS will write an output with predictions to the command prompt and it will generate an image with the predictions in /pro-img/
