MLOS
MLOS copied to clipboard
microsoft
MLOS is a project to enable autotuning for systems.
Would like to be able to execute multiple experiment trials in parallel
Logging this here for now: I've temporarily removed the markdown link lint check from the required checks to pass since it's intermittently failing on sites that look up to me....
We need it to test the logic updated in #412
Further generalizing this via async `telemetry` collection during the process might be nice too.
An optimization we could make, especially for Windows to select pwsh instead of cmd, but perhaps also for Linux to do sh instead of bash or some such (though there...
- [ ] Add a `schema_version` table. - [ ] If it doesn't exist, infer version 0 - [ ] Maintain code that adjusts the tables step by step from...
Per discussions with @motus, one way we could implement an ensemble of optimizers (or LLMs) for next config suggestion could be to model a collection of Optimizers as a CompositeOptimizer....
It is very confusing to have two dicts for `config` and `params` as arguments to many methods in, e.g., `AzureVMService` and `AzureConfigService`. We should be more explicit in naming and...
Sometimes user scripts don't return a `score` value, even though they exit 0 (indicating SUCCESS). In that case we can do a couple of things: - abort immediately in order...