Apologies if this is out of place. I’d like to contribute to PyTorch and looking for a good place to start. I searching for ‘1hr’ tags in current issues but did not find any. Various open issues have a ‘good first issue’ or a ‘bootcamp’ tag but they either have folks working on them or are very old. If anyone could use some help or has any tips please let me know.
5 Likes
Hey!
Did you had a chance to look at our newly written contribution guide: PyTorch Contribution Guide — PyTorch master documentation ?
Don’t hesitate to follow up here if that doesn’t answer your question and give feedback if it is not clear.
1 Like
Hi @albanD, I did - I looked through the issue tracker and couldn’t see anything obvious I’d know how to fix, so I looked for the mentioned tags (e.g., ‘1hr’ and ‘bootcamp’) which led to my post above. I’m sure I’ll find something eventually, but if anyone has any leads I’d be super appreciative. For context I’ve used PyTorch to build out standard neural networks - nothing too fancy.
1 Like
I would also recommend looking at the “actionable” issues: Issues · pytorch/pytorch · GitHub which means that they can be worked on.
You can also add more filters there like module: nn or module: optim if these are the areas you’re more comfortable working in.
2 Likes
Hi @rkrell, I’ve just finished my first PR to pytorch so I’d like to share my experience. I think your direction is correct: the issue with ‘good first issue’ tag is a good start. But since pytorch is a pretty mature project, you may need to take some time to wait for a proper problem. Compared to stable modules, there’re relatively more issues in the emerging modules, such as those released in PT2. You can take a look at them and find one that matches your expertise. I hope this will help.
2 Likes