[lightning] GPU support

[karl-richter]

Adds functionality to train the TimeNet model using an accelerator.
Supports GPU, MPS (M1) automatically and all other available Pytorch Lightning accelerators via manual declaration (eg. TPU etc.). Updated after Lightning merge.

Closes #420
Closes #938

[github-actions]

1efc395

Model Benchmark

Show benchmark results

Benchmark	Metric	main	current	diff
AirPassengers	MAE_val	15.2698	15.2698	0.0%	✅
AirPassengers	RMSE_val	19.4209	19.4209	0.0%	✅
AirPassengers	Loss_val	0.00195	0.00195	0.0%	✅
AirPassengers	RegLoss_val	0	0	0.0%	✅
AirPassengers	epoch	89	89	0.0%	✅
AirPassengers	MAE	9.82902	9.82902	0.0%	✅
AirPassengers	RMSE	11.7005	11.7005	0.0%	✅
AirPassengers	Loss	0.00056	0.00056	0.0%	✅
AirPassengers	RegLoss	0	0	0.0%	✅
PeytonManning	MAE_val	0.64636	0.64636	-0.0%	✅
PeytonManning	RMSE_val	0.79276	0.79276	-0.0%	✅
PeytonManning	Loss_val	0.01494	0.01494	-0.0%	✅
PeytonManning	RegLoss_val	0	0	0.0%	✅
PeytonManning	epoch	37	37	0.0%	✅
PeytonManning	MAE	0.42701	0.42701	0.0%	✅
PeytonManning	RMSE	0.57032	0.57032	-0.0%	✅
PeytonManning	Loss	0.00635	0.00635	0.0%	✅
PeytonManning	RegLoss	0	0	0.0%	✅
YosemiteTemps	MAE_val	1.72949	1.72948	-0.0%	✅
YosemiteTemps	RMSE_val	2.27386	2.27386	-0.0%	✅
YosemiteTemps	Loss_val	0.00096	0.00096	-0.0%	✅
YosemiteTemps	RegLoss_val	0	0	0.0%	✅
YosemiteTemps	epoch	84	84	0.0%	✅
YosemiteTemps	MAE	1.45189	1.45189	-0.0%	✅
YosemiteTemps	RMSE	2.16631	2.16631	-0.0%	✅
YosemiteTemps	Loss	0.00066	0.00066	0.0%	✅
YosemiteTemps	RegLoss	0	0	0.0%	✅

Model Training

PeytonManning

YosemiteTemps

AirPassengers

[codecov-commenter]

Codecov Report

Merging #961 (b8be966) into main (135e8b1) will decrease coverage by 0.25%.
The diff coverage is 13.33%.

@@            Coverage Diff             @@
##             main     #961      +/-   ##
==========================================
- Coverage   87.68%   87.43%   -0.26%     
==========================================
  Files          17       17              
  Lines        4433     4448      +15     
==========================================
+ Hits         3887     3889       +2     
- Misses        546      559      +13     

Impacted Files	Coverage Δ
neuralprophet/utils.py	80.06% <7.14%> (-3.28%)	⬇️
neuralprophet/forecaster.py	87.99% <100.00%> (+0.01%)	⬆️

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

[Kevin-Chen0]

Can you make test case(s) for this to improve the codecov?

[karl-richter]

Can you make test case(s) for this to improve the codecov?

@Kevin-Chen0 Note sure if that makes sense in this case, bc the machine would need to have a GPU / accelerator aside from CPU to test the feature. As far as I know thats not available / reasonable in gh actions.

[Kevin-Chen0]

Can you resolve the merge conflict first?

[Kevin-Chen0]

Karl, can you address the two questions I have before I approve? In the meantime, I'm changing the status back to needs fix. Thx.

[noxan]

Code looks good to me and no merge conflicts.

I agree with @Kevin-Chen0 that it would be great to have test cases, yet as @karl-richter pointed out there is no option with github actions I'd be aware of - if someone knows any good way lets create test cases with a follow up issue 👍

[geneman]

running with CUDA does not work: RuntimeError: Expected all tensors to be on the same device, but found at least two devices, cpu and cuda:0!