Hi Ran,
Do you remember how it was solved before (for kunpeng2) ?
https://anaconda.org/conda-forge/tensorflow does not provide
linux-aarch64 build even for 2.17.0 in the conda-forge channel. Your
package uses 2.10.0.
The anaconda channel does provide it -
https://anaconda.org/anaconda/tensorflow, but AFAIK using this channel is
non-free.
The builder setup docs list tensorflow as a PIP dependency:
https://github.com/Bioconductor/BBS/blob/devel/Ubuntu-files/22.04/pip_pkgs.txt#L8
biocbuild at kunpeng2 ~> pip install tensorflow
Defaulting to user installation because normal site-packages is not
writeable
Requirement already satisfied: tensorflow in
/usr/local/lib/python3.9/site-packages (2.12.0)
Requirement already satisfied: tensorflow-cpu-aws==2.12.0 in
/usr/local/lib64/python3.9/site-packages (from tensorflow) (2.12.0)
I.e. tensorflow 2.12.0 is installed via pip, but conda does not use pip
packages.
Regards,
Martin
On Fri, Aug 16, 2024 at 4:51?AM RAN HU <huran at ucla.edu> wrote:
Hi,
I am the maintainer of the cfTools package. I found cfTools have build
errors on palomino7/8 and kunpeng2 for BioC 3.19 and 3.20, which appear to
be caused by an environment issue related to a Python dependency. This has
happened before and was solved. Could you help me solve this again? And is
there a way to avoid this error in future when there is a new release?
Thank you very much!
Best,
Ran
[[alternative HTML version deleted]]