Re: alphafold Debian packaging ?
On 2022-01-12 17:34, Gard Spreemann wrote:
> Andrius Merkys <merkys@debian.org> writes:
>> On 2022-01-11 16:14, Gard Spreemann wrote:
>>> I'm taking the liberty to forward this to the debian-ai list as well.
>>
>> Thanks. Highly relevant issue for debian-ai@ might be the retraining of
>> the neural network model.
>
> Indeed! I was thinking that it might be an interesting – and
> complicated! – test case for Mo Zhou's draft ML policy [2].
>
> [2] https://salsa.debian.org/deeplearning-team/ml-policy
Sure!
>> In addition to being pre-trained, the model itself is licensed under
>> CC-BY-NC-4.0, thus not suitable for Debian main. I could not find a
>> guide for retraining, and an issue on their tracker [1] asks for the
>> same.
>
> Hmm, interesting! Where did you see this? Regarding the output
> predictions, alphafold.com states CC-BY-4.0 (in two different places):
>
>> All of the data provided is freely available for both academic and commercial use under Creative Commons Attribution 4.0 (CC-BY 4.0) license terms. If you use this resource please cite the following papers:
>> […]
>
> And their code repository is Apache. Or did you find the actual
> pretrained models somewhere under CC-BY-NC?
Interesting. Maybe I am looking at some other source. Am I right to
assume we are talking about [3]? If so, it says that the parameters are
CC-BY-NC here: [4].
[3] https://github.com/deepmind/alphafold
[4] https://github.com/deepmind/alphafold#model-parameters
Best,
Andrius
Reply to: