vsha96 / mllib

Machine Learning in Haskell
BSD 3-Clause "New" or "Revised" License
29 stars 1 forks source link

Initial performance tests should be done to get an idea whether currest containers are fast enough. Probably a fast (mutable) container should be developed with basic linear algebra operations on it. #10

Open s1m0000n opened 2 years ago

s1m0000n commented 2 years ago

Is your feature request related to a problem? Please describe. I think Haskell's lists are not the most performant solution. I guess there could be speed / memory consumption issues when training & inferencing on reasonable to life-usage datasets.

Describe the solution you'd like

vsha96 commented 2 years ago

I'm glad you noticed this. I've been thinking about it from the very beginning, but my tests were primitive (I used TimeIt).

For further performance testing, we should use criterion.