| About |
| Downloads |
| Contributions |
| Documentation |
| FAQ |
| Contacts |
| Known Bugs |
SPArse Modeling Software
now open-source!
Known Bugs and Solutions
Bug 1 (unsolved): On some computer, the toolbox complains about the missing file libatlas.so.3gf
Next minor update will include this file. In the meantime, you should use the version mkl32 if you have this problem.
Bug 2 (solved in version >= 1.01): When I forget to pass param.eps to mexOMP, Matlab crashes.
There is a bug in the way I protected the mex files when some input parameters are missing. The version 1.01 take care of this problem.
Bug 3 (unsolved): mexTrainDL_Memory does not perform as well as mexTrainDL when param.lambda is very small.
mexTrainDL uses LARS to solve l1-decomposition problems. mexTrainDL_Memory uses coordinate descent.
When the regularization parameter is very small, coordinate descent can be very slow and so mexTrainDL_Memory becomes unusable.
This is a particular case where mexTrainDL should be used instead.
Bug 4 (unsolved): mexConjGrad does only work with 5 arguments.
This will be solved in the next release.
Bug 5 (unsolved): When param.mode=1, and when all elements of the dictionary are used for the
reconstruction of a signal (due to a very small regularization), mexLasso does not return the right solution.
This will be solved in the next release.
Bug 6 (unsolved): When the l2-norm of my input vectors is very big or very small, mexLasso
does not return the correct solution.
This is a numerical stability problem that will be solved in the next release. In the meantime, you should rescale your data.