Changes in version 1.0.2 (2026-04-10)                  

  - Added - Added a new function adjust.model for aligning two LCA or
    LPA objects.
  - Added - Added visualization plots (S3 method plot) for LCA and LPA.
  - Fixed - Resolved the case sensitivity issue of observed indicator
    names when calling Mplus
  - Fixed - Improved Python environment selection in the
    install_python_dependencies function.
  - Fixed - Fixed the description issues in the comment documentation.
  - Fixed - Enhanced the stability of the LRT.test.Bootstrap function.
  - Fixed - Removed unnecessary Mplus statements from LCA and LPA.
  - Fixed - Improved the efficiency of Python code calling NVIDIA GPUs
    (Cuda + cuDNN + Pytorch) for deep network training.

                 Changes in version 1.0.1 (2026-02-27)                  

  - Fixed - Resolved the configuration issue with control.NNE.
  - Fiexd - Resolved return parameter res$params$par naming issue in
    LCA.
  - Added - For LCA, LPA, LCPA, and LTA, users can control whether to
    enable the self-attention mechanism (i.e., transformer encoder)
    through the use.attention arguments.

                 Changes in version 1.0.0 (2026-01-22)                  

  - Initial release