| initialization helps :) woot! |
|
More...
|
about 13 years ago
|
| fixed scaling of primary objective |
|
More...
|
about 13 years ago
|
| cleanup after fix of primal objective computation |
|
More...
|
about 13 years ago
|
| plot fix in latens ssvm |
|
More...
|
about 13 years ago
|
| some more debugging stuff. it only makes sense to compute the primal objective after optimization ^^ and lacks are always positive obv.... d'oh. |
|
More...
|
about 13 years ago
|
| hacky regularization of qp in ssvm. |
|
More...
|
about 13 years ago
|
| cleanup in latent svm, better plotting, parallelization, random initialization. |
|
More...
|
about 13 years ago
|
| encouraging results. reproducable? |
|
More...
|
about 13 years ago
|
| parallizing inference with joblib |
|
More...
|
about 13 years ago
|
| Fix in latent example plotting |
|
More...
|
about 13 years ago
|
| FIX in StupidLatentSVM |
|
More...
|
about 13 years ago
|
| rename multinomial grid example. |
|
More...
|
about 13 years ago
|
| toy dataset: latent crosses |
|
More...
|
about 13 years ago
|
| some fixes in latent structured svm |
|
More...
|
about 13 years ago
|
| cleanup of latent crf example |
|
More...
|
about 13 years ago
|
| more robust inference with libdai (normalization of parameters) |
|
More...
|
about 13 years ago
|
| exhaustive test for loss-augmented inference in latent variable model |
|
More...
|
about 13 years ago
|
| enh inference method as option to crf |
|
More...
|
about 13 years ago
|
| more tests for latent svm, fixed minor bugs |
|
More...
|
about 13 years ago
|
| fix in latent crf, added test for latent crf. |
|
More...
|
about 13 years ago
|
| fix tests for new dataset, format weights to make clear what they mean. |
|
More...
|
about 13 years ago
|
| modified binary datasets for easier testing in latent variable cases |
|
More...
|
about 13 years ago
|
| rename MultinomialGridMRF to GridMRF |
|
More...
|
about 13 years ago
|
| Remove binary grid CRF |
|
More...
|
about 13 years ago
|
| some more experiments... |
|
More...
|
about 13 years ago
|
| generate crossed explicit toy data so that the unaries don't know about centers. |
|
More...
|
about 13 years ago
|
| tests for multinomial crfs |
|
More...
|
about 13 years ago
|
| starting on latent grid crf |
|
More...
|
about 13 years ago
|
| switch back to qpbo for running some tests faster |
|
More...
|
about 13 years ago
|
| make smaller "easy" dataset, make binary "crosses" dataset |
|
More...
|
about 13 years ago
|
