Open edjez opened 9 months ago
Hey @edjez, there is another fork of this repo which has a number of bugfixes and additional tools to resolve exactly what you are looking for. Have a look at the README here (ctrl+f for "combine_logs"): https://github.com/michaeltremeer/azure-openai-benchmark. Note that you will need to rerun your tests using the fork, since right now none of the run info is logged in this official repo
There is probably no need for this as you can pipe stdout to jq
and reformat jsonl any which way you want.
Customers are telling us they have to write custom code today to output to a nicely formatted CSV to analyze results on their own excel/google sheets/