[Error: All translation engines failed for batch: MADLAD batch translation failed: CUDA out of memory. Tried to allocate 2.00 MiB. GPU 0 has a total capacity of 23.87 GiB of which 3.62 MiB is free. Process 3280094 has 228.00 MiB memory in use. Process 2050901 has 248.00 MiB memory in use. Process 3668120 has 1.09 GiB memory in use. Process 3668111 has 1.09 GiB memory in use. Process 3668113 has 1.88 GiB memory in use. Process 3668109 has 1.12 GiB memory in use. Process 3668117 has 1.50 GiB memory in use. Process 3668112 has 1.42 GiB memory in use. Process 3668118 has 1.86 GiB memory in use. Process 3668110 has 1.15 GiB memory in use. Process 3693685 has 1004.00 MiB memory in use. Process 3668125 has 760.00 MiB memory in use. Process 3736374 has 1.19 GiB memory in use. Including non-PyTorch memory, this process has 9.36 GiB memory in use. Of the allocated memory 9.12 GiB is allocated by PyTorch, and 74.92 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)]