search

automl / Auto-PyTorch

Automatic architecture search and hyperparameter optimization for PyTorch
Apache License 2.0
2.37k stars 287 forks source link

when there is no training at all, i think it would be the best to raise Error #63

Closed maxmarketit closed 3 years ago

maxmarketit commented 4 years ago

During training, I somehow end up with these messages,

00:12:08 Cannot re-initialize CUDA in forked subprocess. To use CUDA with multiprocessing, you must use the 'spawn' start method Process pynisher function call:

And most of the time, I get nothing from training for 5 min, 10 mins, even days.

So i guess i would be better to raise Error and stop the training if the problem is not solvable by autoPyTorch.

ravinkohli commented 3 years ago

Hey, thanks for this question. Actually, we have been working on a new version of autopytorch that is available on the branch `development, for docs, see this. Now we feel it's ready for people to test their workflows with this version.

So I suggest you try out your data with the development version. For an example on how to use the API, find it here

Feel free to ask any questions about the implementation here. And if you feel the issue is resolved, kindly close it.