Re: alphafold Debian packaging ?
On 12.01.22 17:24, M. Zhou wrote:
Even more complicated is the underlying software dependency tree.
alphafold depends on dm-haiku, jax, tensorflow.
dm-haiku depends on jax.
jax depends on XLA from tensorflow.
tensorflow still in NEW.
Long way to go. Mhhh.
That is what I had thought, too. On
https://docs.google.com/spreadsheets/d/1tApLhVqxRZ2VOuMH_aPUgFENQJfbLlB_PFH_Ah_q7hM/edit#gid=1840067013
I once collected the immediate dependencies and can confirm that it
there is quite a bit of homework out there. If the only purpose is to
bring AlphaFold to Debian then I suggest not to do it. The benefit may
be to improve the software basis of Debian at large with it. But also, I
can do something good elsewhere and all these releases update so
quickly, to me it stopped being fun.
What's also complicated is the GPU support. Currently the only
working modern deep learning framework in our archive is pytorch,
which is only compiled with cpu support.
pytorch-cuda requires cudnn. I gave up cudnn packaging a few
times and I eventually realized that I dislike working on
nvidia stuff even if I have to use it.
pytorch-rocm is a good way to go. As you can see on debian-ai@
people are still working hard to get ROCm into debian.
Intel GPU support is too new to evaluate.
To have ROCm as a source package with Debian will be good for everyone.
IMHO this may actually dwarf the benefit of having AlphaFold with us.
So, yes, if something like AlphaFold is something that we need as a
carrot to improve our Java infrastructure, then I am all up for it.
The prospect to have a community-regenerated knowledge base for
AlphaFold is also very tempting.
Steffen
On Wed, 2022-01-12 at 16:54 +0100, Gard Spreemann wrote:
Andrius Merkys <merkys@debian.org> writes:
On 2022-01-12 17:34, Gard Spreemann wrote:
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
Interesting indeed! So we have:
– Training data: A plethora of different licenses.
– Code: Apache
– Trained model: CC-BY-NC-4.0
– Output of said trained model: CC-BY-4.0 [5]
Nightmarish!
[5] See under "license and attributions" on https://alphafold.com
-- Gard
Reply to: