/u/cboul/.local/lib/python3.10/site-packages/trl/trainer/sft_trainer.py:294: UserWarning: You passed a tokenizer with `padding_side` not equal to `right` to the SFTTrainer. This might lead to some unexpected behaviour due to overflow issues when training a model in half-precision. You might consider adding `tokenizer.padding_side = 'right'` to your code.
OutOfMemoryError: CUDA out of memory. Tried to allocate 30.00 MiB. GPU 0 has a total capacity of 23.67 GiB of which 38.19 MiB is free. Process 36053 has 11.64 GiB memory in use. Process 31527 has 7.32 GiB memory in use. Including non-PyTorch memory, this process has 4.63 GiB memory in use. Of the allocated memory 4.34 GiB is allocated by PyTorch, and 9.73 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)