Make PyTorch models up to 40% faster! Thunder is a source to source compiler for PyTorch. It enables using different hardware executors at once; across one or thousands of GPUs.
triage review: let's do it! fdm is no longer supported; let's do what we need to refactor its use in tests or copy/fork the relevant code and update it
It seems like it is quite easy to make Thunder support NumPy 2, modulo dependencies (
fdm
, for example).A temporary fix could be disabling envs with NumPy 2 in our CIs, https://github.com/Lightning-AI/lightning-thunder/pull/617.