Erik Bernhardsson
Erik Bernhardsson
i'm definitely open to adding this provided it can use as many standard datasets and algorithms as possible, and it can be reasonably concise. https://github.com/erikbern/ann-benchmarks/blob/master/ann_benchmarks/datasets.py is where all the datasets...
It optimizes for storage space: _Even though there are ANN libraries producing reasonable recall in less than 1 ms, for our use case we can accept sacrificing some of that...
@chasingegg curious if you can elaborate. what's different with the scenario they are trying to optimize for?
@zsz00 i would love to add it. maybe the authors can take a look. @una-dinosauria ?
if my memory serves me right, i believe you can just do ``` python install.py --algorithm lopq python run.py --algorithm lopq ``` does that work? we can add it to...
sure – want to give it a shot?
sure – want to go ahead and give it a shot?
i don't think any of them do, unfortunately :(
I don't have any great recommendations, but your idea makes sense though. You could even formalize it and have indexes that have size even powers of two, and any time...
@yurymalkov yeah i've never heard of it, but think that's what i meant. basically play 2048 with the indexes :)