Search Criteria
Package Details: python-torchmetrics 1.4.1-1
Package Actions
Git Clone URL: | https://aur.archlinux.org/python-torchmetrics.git (read-only, click to copy) |
---|---|
Package Base: | python-torchmetrics |
Description: | Machine learning metrics for distributed, scalable PyTorch applications |
Upstream URL: | https://github.com/PyTorchLightning/metrics |
Licenses: | Apache-2.0 |
Submitter: | hottea |
Maintainer: | alhirzel |
Last Packager: | alhirzel |
Votes: | 2 |
Popularity: | 0.014387 |
First Submitted: | 2021-04-01 09:02 (UTC) |
Last Updated: | 2024-08-09 02:16 (UTC) |
Dependencies (11)
- python-lightning-utilitiesAUR
- python-nltk
- python-regex (python-regex-gitAUR)
- python-sentencepieceAUR (python-sentencepiece-gitAUR)
- python-torchvision (python-torchvision-gitAUR, python-torchvision-rocm-binAUR, python-torchvision-rocmAUR, python-torchvision-cuda)
- python-tqdm
- python-build (make)
- python-installer (python-installer-gitAUR) (make)
- python-fidelity (optional)
- python-lpipsAUR (optional)
- python-transformersAUR (optional)
Latest Comments
Romanovzky commented on 2024-04-20 15:45 (UTC)
Why does this depend on
python-transformers
? It ends up installing a lot of things that are completely unnecessary for PyTorch lightning as this is set as dependency ofpython-pytorch-lightning
. See here for strict dependencies: https://github.com/Lightning-AI/torchmetrics/blob/master/requirements/base.txtFreed commented on 2023-03-11 03:40 (UTC) (edited on 2023-03-11 03:43 (UTC) by Freed)
@alhirzel I have invited you in python-torch-fidelity and python-lpips. In
/usr/lib/python3.10/site-packages/torchmetrics-0.11.1-py3.10.egg-info/requires.txt
I found more packagesscipy
, ... it looks like some of them are optdepends, too. (scipy is the optdepend for FID)alhirzel commented on 2023-03-10 17:41 (UTC) (edited on 2023-03-10 17:43 (UTC) by alhirzel)
@Freed I bumped this package, thanks for your note. Do you want me to package FID on the AUR?
Freed commented on 2023-03-10 17:14 (UTC)
Miss optdepends: python-lpips for LPIPS and python-fidelity for FID.