v2.1.0

What's Changed

• Add new particle samplers by @austin-hoover in #26

Full Changelog: v2.0.0...v2.1.0

v2.0.0

Switched to PyTorch (instead of NumPy) to enable gradient-based sampling algorithms and running on GPU.

What's Changed

• Switch to PyTorch by @austin-hoover in #24
• Edit README.md by @austin-hoover in #25

Full Changelog: v1.1.7...v2.0.0

v1.1.7

What's Changed

• Do not store ipynb outputs on git by @austin-hoover in #20
• Simplify sampling example to 2D by @austin-hoover in #22
• Add pre-commit hooks by @austin-hoover in #23

Full Changelog: v1.1.6...v1.1.7

v1.1.6

What's Changed

• Organize examples by @austin-hoover in #11
• Add HDR reconstruction example by @austin-hoover in #13

Full Changelog: v1.1.5...v1.1.6

v1.1.5

Small changes to examples.

Full Changelog: v1.1.4...v1.1.5

v1.1.4

• Add 6D reconstruction example from paper "N-dimensional maximum-entropy tomography via particle sampling".
• Add traditional 2D CT example with comparisons to FBP and SART on high-resolution images.
• Automatically normalize projections passed to MENT.
• Fix histogram fractional threshold.
• Enforce learning rate between 0 and 1 in Gauss-Seidel updates
• Create new integration method in which the density is evaluated on the entire N-dimensional grid and summed over the integration axes. This is a faster method that can work well in 2D/3D. The method can be turned on by setting integration_loop=False in ment.MENT class.

Full Changelog: v1.1.3...v1.1.4

v1.1.3

What's Changed

• Example 4D nonlinear ring example in #8
• Add 4D SNS example in #9
• Improved covariance matrix fitter
• Improved Metropolis Hastings sampler debugger

Full Changelog: v1.1.2...v1.1.3

v1.1.2

• Improve covariance matrix fitting routines in #5

v1.1.1

Fix bug when initializing MENT with no arguments.

v1.1.0

• Added new prior UniformSphericalPrior and added truncate option to Gaussian prior.
• Cleaned up MCMC sampler, added noise option, added documentation
• Added sample, and cov methods to HistogramND class.
• Added sample, and std methods to HistogramND class.
• Added CovFitter class to fit ND covariance matrices to measurements using evolutionary optimization.
• Added longitudinal phase space tomography example using PyORBIT.
• Fixed a few bugs.

Full Changelog: v1.0.2...v1.1.0