opentensor/async-substrate-interface

TODO: Preload the caches

Closed this issue · 1 comments

The LRU caches should be able to be preloaded at startup. Perhaps we should look into dumping this info at the close of the object, and attempting to reload from said cache at startup.

We can do this with sync, but won't work for async https://joblib.readthedocs.io/en/latest/memory.html