Mot

Latest version: v0.11.4

Safety actively analyzes 681866 Python packages for vulnerabilities to keep your Python projects secure.

Scan your dependencies

Page 4 of 11

0.7.1

===================

Added
-----
- Adds private memory to the list of KernelData classes. This makes it possible to add private memory elements to the kernel data.

Other
-----
- Small update to the twalk sampler.

0.7.0

====================
By removing all local variable out of non-kernel functions, this version should now be compatible with POCL (tested with version 1.1).

Changed
-------
- Updates to the samplers. Adds initial t-walk sampler implementation.

Other
-----
- Removed warning filter in the CL function.
- Refactored the constructor of the cl function parameter.
- Removed the CL load balancing. It was not very useful.
- Modified the Nelder-Mead simplex default scales for slightly better fits.

0.6.14

====================

Added
-----
- Adds linear cubic interpolation method for interpolating on a grid.
- Adds a CL function for Simpson's rule numerical integration

Fixed
-----
- Bugfix in the CL multi-functions parser.

Changed
-------
- Work on moving local variable declarations outside of non-kernel functions. This should in the future allow running MOT on LLVM OpenCL implementations. More work to be done.
- Improved the default settings of the Subplex optimizer.

0.6.13

====================

Added
-----
- Adds the logpdf for the Gamma and the Normal distribution.
- Adds an example of fitting a Gamma distribution.

Changed
-------
- Changed the function signature of the legendre polynomial computations.

0.6.12

====================

Fixed
-----
- Out of bounds fix in the legendre polynomial computations.

0.6.11

====================

Changed
-------
- Updates to the SCAM MCMC routines default settings.

Other
-----
- Removed CL extra from the CL functions and added support for CLCodeObjects for injecting simple CL code as a dependency.
- Small local memory optimization to the LM optimization routine.

Page 4 of 11

© 2024 Safety CLI Cybersecurity Inc. All Rights Reserved.