5 Commits

Author	SHA1	Message	Date
based	088814edd6	prioritize non filtered models when coming across the same model tier in azure ... also reverted the azure checker to being multithreaded instead of asynchronous, the async version does nothing but choke on large key counts even with a small semaphore value. if theres any sort of rate limit in place on something then async is slop. run_in_executor was also causing timeout errors and ignoring caught exceptions so it's been removed and replaced with the old as_completed method. it gets passed to an executor in the end either way so there shouldn't be much of a performance difference, if any.	2024-01-27 19:21:52 +10:00
based	80bea77242	clean up requirements.txt	2024-01-18 10:29:49 +10:00
based	22de3b0754	added async support for various services	2024-01-18 09:14:54 +10:00
based	266bc976ea	Added Google Vertex AI support	2023-12-11 03:44:49 +10:00
based	cf87885b5e	updated README.md	2023-12-06 13:01:59 +10:00