Make rankme work with lightning

[dungscout96]

Currently we're receiving error when running rankme.compute()

  File "/home/dung/eeg-ssl/libs/evaluation.py", line 31, in compute
    _, S, _ = torch.linalg.svd(embs)
RuntimeError: cusolver error: CUSOLVER_STATUS_INVALID_VALUE, when calling `cusolverDnSgesvdj_bufferSize(handle, jobz, econ, m, n, A, lda, S, U, ldu, V, ldv, lwork, params)`. If you keep seeing this error, you may use `torch.backends.cuda.preferred_linalg_library()` to try linear algebra operators with other supported backends. See https://pytorch.org/docs/stable/backends.html#torch.backends.cuda.preferred_linalg_library

[dungscout96]

this seems to only break with large batch size. Pattern for initializing and running multiple evaluators is implemented in ffc775d