Open Dormiveglia-elf opened 1 week ago
I have reduced the number of timeout error by adjusting the max_workers to 60 and time_out parameter to 120 mentioned in https://docs.ragas.io/en/stable/howtos/customizations/_run_config/?h=time
But still some error existing, I think 120 is a very long time. Any ideas?
max_workers
Further reducing max_workers solved the same issue for me. I set it to be 8 for my locally hosted models.
hey anothe approach is batch_size
argument for evalaute()
https://docs.ragas.io/en/stable/references/evaluate/#ragas.evaluation.evaluate
something else you could try? easier to reason about IMO
ERROR:ragas.executor:Exception raised in Job[2]: TimeoutError() ERROR:ragas.executor:Exception raised in Job[7]: TimeoutError() ERROR:ragas.executor:Exception raised in Job[10]: TimeoutError() ERROR:ragas.executor:Exception raised in Job[15]: TimeoutError() ....
I use GPT-4o and encountered many timeout error, any ideas?