build: upgrade torch and let transformers pick the version

[julian-risch]

Related Issues

• fixes Upgrade torch to 1.13.0 #3683
• related issue Better model security with new PyTorch version #3521

Proposed Changes:

• PyTorch 1.13.1 was released today and CI test passed. Version 1.13.0 and 1.13.1 both work now with Haystack so it is save to upgrade. Let's upgrade! We will use the same version range as transformers so we let transformers install torch as one of it's dependencies.

How did you test it?

CI and so manual tests

Notes for the reviewer

Checklist

• I have read the contributors guidelines and the code of conduct
• I have updated the related issue with new insights and changes
• I added tests that demonstrate the correct behavior of the change
• I've used the conventional commit convention for my PR title
• I documented my code
• I ran pre-commit hooks and fixed any issue

[vblagoje]

@julian-risch This is great! So we can set the torch to "torch>=1.7,!=1.12.0" and move on?

[julian-risch]

I think we need to exclude 1.13.0 too. Let's double check and have a quick chat.

[julian-risch]

To sum up the discussion: we let transformers handle the pytorch version, which results in "torch>=1.7,!=1.12.0", as of now: https://github.com/huggingface/transformers/blob/main/setup.py#L169

[vblagoje]

@mayankjobanputra can you approve this one, if all is good?

[mayankjobanputra]

I also checked the logs... Looks super perfect!

Requirement already satisfied: torch>=1.3.0 in /opt/hostedtoolcache/Python/3.7.15/x64/lib/python3.7/site-packages (from espnet==202209->farm-haystack==1.13.0rc0) (1.13.1)