FletcherAtFADU

Checkpointing failure with llama_7b_zero3 workload: 'NoneType' object has no attribute 'save_checkpoint'

1

Hi, I'm trying to benchmark checkpoint performance using the `llama_7b_zero3` workload in DLIO. I'm running the benchmark with the following command: ```bash mpirun -np 16 dlio_benchmark workload=llama_7b_zero3 ++workload.workflow.generate_data=True ++workload.workflow.checkpoint=True ```...