Closed csukuangfj closed 2 years ago
Hey @csukuangfj :wave:
Thanks for your detailed report, I've set the project limit for k2
to 20 GB on PyPI. Please be mindful of the frequency of releases at that size, you might want to explore the Stable C ABI to at least have 1 wheel per platform, rather than 1 per python version https://docs.python.org/3/c-api/stable.html
Have a nice week!
Project URL
https://pypi.org/project/k2
Does this project already exist?
New limit
20 GB
Update issue title
Which indexes
PyPI
About the project
k2, the next-gen Kaldi, is to be able to seamlessly integrate Finite State Automaton (FSA) and Finite State Transducer (FST) algorithms into autograd-based machine learning toolkits like PyTorch and TensorFlow.
We started the project in 2020 and have been actively developing it. You can find from https://github.com/k2-fsa/k2/graphs/contributors that there are commits very week.
How large is each release?
290 MB total
4 Linux wheels (one for each Python version, 3.7, 3.8, 3.9, 3.10), each wheel is ~70 MB since there are CUDA code in the project and we have to compile it for every GPU architecture. 4 macOS wheel; each wheel is ~3 MB
How frequently do you make a release?
Roughly twice per month
Code of Conduct