fix incomplete RunningMean

[Borda]

What does this PR do?

Until the mean is once rotated it returns random numbers.

t = torch.Tensor(5)
Out[5]: tensor([3.8516e-37, 0.0000e+00, 1.6114e-19, 1.8475e+31, 3.8715e-37])
t[0] = 1
Out[7]: tensor([1.0000e+00, 0.0000e+00, 1.6114e-19, 1.8475e+31, 3.8715e-37])
t.mean()
Out[8]: tensor(3.6949e+30)

PR review

Anyone in the community is free to review the PR once the tests have passed.
If we didn't discuss your PR in Github issues there's a high chance it will not be merged.

Did you have fun?

Make sure you had fun coding 🙃

[Borda]

@williamFalcon why you changed it to return self.memory.mean()?
if you have [0, 0, 0, 0, 0] and place single value 1, then you get 0.2 as mean instead of 1

[pep8speaks]

Hello @Borda! Thanks for updating this PR.

• In the file pytorch_lightning/core/lightning.py:

Line 1529:111: E501 line too long (111 > 110 characters)

Comment last updated at 2020-03-30 21:34:42 UTC