Nevergrad

Latest version: v1.0.2

Safety actively analyzes 613631 Python packages for vulnerabilities to keep your Python projects secure.

Scan your dependencies

Page 1 of 4

0.11.0

- additional benchmarks
- adding tests in the CI
- helpers for choosing algorithms for different problems

0.10.0

- fix the issue with the Boston dataset

0.9.0

- Add many quasi-opposite variants
- Improve the Dagsduhloid benchmark
- Get rid of legacy pypi solver in CircleCI
- Update Pypi in CircleCI
- Switch to Python 3.8

0.8.0

- Add the Dagstuhloid benchmark
- Add yet another group of metamodels

0.7.0

- Fix links
- Add metamodels
- Update for weighted multiobjective optimization with differential evolution

0.6.0

Breaking changes

- Removed `descriptor` field of parameters which had been deprecated in previous versions. Use `function` field instead to specify if the function
is deterministic or not [X](https://github.com/facebookresearch/nevergrad/pull/X).

Important changes

- `TransitionChoice` behavior has been changed to use bins instead of a full float representation. This may lead to slight
changes during optimizations. It can also be set as unordered for use with discrete 1+1 optimizers (experimental)
- Adding NGOptRW, presumably better than NGOpt for real-world problems.
- Making some dependencies optional because running was becoming too complicated.
- Adding the NLOPT library.
- Adding smoothness operators for discrete optimization.

Other changes
- Adding YAPBBOB, with a parameter regulating YABBOB-like problems so that the distribution of the optimum is less rotationally invariant.
- Adding constrained counterparts of YABBOB: yapenbbob (a few constraints), yaonepenbbob (single constraint), yamegapenbbob (many constraints).
- Improvements in the photonics benchmarks.
- Externalizing CompilerGym.
- Making some tests less flaky.
- Adding Simulated annealing and Tabu search.
- Making the code more robust to Gym environments.

Page 1 of 4

© 2024 Safety CLI Cybersecurity Inc. All Rights Reserved.