Comparing
optimizations
(561753b
) withmain
(707eacf
)+1%
- Improvements
- 0
- Regressions
- 0
- Untouched
- 5
- New
- 0
- Dropped
- 0
- Ignored
- 0
Benchmarks
Passed
open benchmark
benches/open_benchmark.rs::benches::open_data_benchmark::open benchmark+2%
81.6 ms
79.8 ms
K-Matrix NLL Performance[3]
benches/kmatrix_benchmark.rs::benches::kmatrix_nll_benchmark::K-Matrix NLL Performance[3]+1%
43.9 ms
43.6 ms
K-Matrix NLL Performance[4]
benches/kmatrix_benchmark.rs::benches::kmatrix_nll_benchmark::K-Matrix NLL Performance[4]+1%
43.9 ms
43.6 ms
K-Matrix NLL Performance[1]
benches/kmatrix_benchmark.rs::benches::kmatrix_nll_benchmark::K-Matrix NLL Performance[1]+1%
43.9 ms
43.6 ms
K-Matrix NLL Performance[2]
benches/kmatrix_benchmark.rs::benches::kmatrix_nll_benchmark::K-Matrix NLL Performance[2]+1%
43.9 ms
43.6 ms
Commits
Click on a commit to change the comparison rangeBase
main
707eacf
+1%
feat: add intermediate `Model` and serialization updates
Big commit which has BREAKING CHANGES! This adds an intermediate struct called `Model` which has `serde` implementations. `Model`s can be saved and loaded and completely contain everything needed to make an `Evaluator` or `NLL` except for the `Dataset`.
To do this, `Amplitude`s need to be (de)serializable. This is a bit tricky, but the `typetag` crate does the trick and it isn't too much of a burden to throw `#[typetag::serde]` on `Amplitude` impls.
561753b
16 days ago
by denehoffman