Using recent libtensorflow and libtorch on machines without AVX2

This actually happens when using LCG_98python3 on a significant part of the CP3 ingrid cluster (ingrid-ui1 included), see the discussion in https://gitlab.cern.ch/cp3-cms/bamboo/-/merge_requests/20#note_3703146 ; I also asked for input/workarounds in https://sft.its.cern.ch/jira/browse/SPI-1691.

Any ideas or suggestions are welcome

To upload designs, you'll need to enable LFS and have an admin enable hashed storage. More information