-
Notifications
You must be signed in to change notification settings - Fork 3.5k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
IterableDataset breaks 1.1 compatibility #491
Comments
good catch. mind submitting a PR? was not the intention of that fix :) We want to support 1.1 |
I have played with versions and it seems to me that it does not run even with 1.2.0; see #462 |
weird. let’s patch it and make a nee release? in the meantime, put a note on the readme to use 0.5.2.1 if you need pt 1.1 or 1.2 support? |
you may rather change requirements.txt and after fix put it back... |
Is this already fixed? I see that torch 1.2.0 is already in requirements.txt ? |
I think not yet...8-) |
Do you suggest torch 1.3.0 ? |
we need to support 1.2.0 :) |
@Borda did we get this fixed? is PT 1.2 supported now? If not, what do we need to do? |
I have not worked on it yet (bit busy with docs), hopefully this week... |
A recently introduced feature unfortunately breaks compability with Pytorch 1.1.0.
Describe the bug
IterableDataset support, introduced in issue 323, requires Pytorch 1.2.0+.
To Reproduce
In a python environment with Pytorch 1.1.0 do:
import pytorch_lightning
Expected behavior
Compatibility with Pytorch 1.1.0. I'm filing it as a bug report rather than a docs fix since the dependency on 1.2.0+ introduced by issue 323 doesn't seem to be intentional.
The text was updated successfully, but these errors were encountered: