Unsupervised-Multi-hop-QA
Unsupervised-Multi-hop-QA copied to clipboard
how can i modify batch_size
error:cuda oom It seems that there is no optional named batch_size in config.if i want to decrease it,what should i do.
I ve got several “batch_size”in the project, such as “train_batch_size”,“ eval_batch_size” and so on. Which one should i decrease to let the project run without "cuda oom".