I'm in the process of updating the tensorflow-example to use TensorFlow 2.6.0 and this adds parity between the code presented in the docs and the example.
There are two minor differences:
1) The tensorflow-example extracts normalizing the pixel values from 0-255 to 0.0-1.0 into a preprocessing step.
2) The tensorflow-example encapsulates all scripts into a
def main():
...
if __name__ == '__main__':
main()
While following this specific best practice might be a bit of an overkill in the example, I feel it definitely would be in the docs.
orasimus
commented
2 years ago
…low-example
I'm in the process of updating the
tensorflow-exampleto use TensorFlow 2.6.0 and this adds parity between the code presented in the docs and the example.There are two minor differences: 1) The
tensorflow-exampleextracts normalizing the pixel values from 0-255 to 0.0-1.0 into a preprocessing step.2) The
tensorflow-exampleencapsulates all scripts into aWhile following this specific best practice might be a bit of an overkill in the example, I feel it definitely would be in the docs.