The Python-library numpy 1.25, which is widely used for scientific computations with multidimensional massifs and matrices, has announced a new release. In addition to containing a large collection of functions with the implementation of various algorithms, numpy is continuously updated to improve the speed of execution and promote new infrastructure.
In this new version, there have been several significant changes:
- Numpy developers have continued to improve the processing and promotion of new infrastructure for the class dtype. This is an essential feature of numpy, and the improvement will make it easier for users to work with data types.
- The speed of execution has increased, making numpy faster and more efficient.
- Prepare for the future release of Numpy 2.0.0 by discontinuing support for some outdated capabilities. Numpy 1.26.0 will be formed after reaching Python 3.12 of the stage of the candidate for release, after which Numpy 2.0.0.
- The formation of Wheel packets based on the standard SIBLITECTION MUSL so that the numpy package can be installed seamlessly on all supported operating systems
- Added support for the Fujitsu C/C++ compiler, expanding the range of compiler options available to numpy users.
- Support for arrays of objects in the einsum function and added support for matrix multiplication at the place using the operator “@=”. These new features make it easier for users to perform complex calculations and work with data more effectively.
The numpy project code is written in the Python language using optimizations in the language and is distributed under the BSD license. To learn more about numpy and download this latest release click here.
/Reports, release notes, official announcements.