Skip to content

Update Main#144

Open
vincent-maillou wants to merge 154 commits into
mainfrom
dev
Open

Update Main#144
vincent-maillou wants to merge 154 commits into
mainfrom
dev

Conversation

@vincent-maillou

Copy link
Copy Markdown
Collaborator

Push all latest development from the new_stats_capa on main.

- gst_small model is now broken, need to check why the internal/external theta doesn't match with other ST models...
-> Bug arose in numpy/cupy 2.4
-> Bug arose in numpy/cupy 2.4
- added saving option for precision matrices
- Implemented memory improvement in the selected inversion using scipy.sparse-like solver
…ties

This PR address the addition of new statistical capabilities to DALIA. Most of them related to post-processing steps (i.e. after the optimization), but also contains new statistical models, new solver capabilities, and new statistical modeling capabilities.

Detailed comments:
- This PR makes a fundamental structural difference in the way the internal and external scales for the hyper-parameters are handled. In more details, the Model() classes have getters and setters that automatically (implictly) perform the conversion.
- This PR also adds some integration testing. These integration tests are not exhaustive of DALIA's modelign capabilities but should catch major failures n code modification.

Notes :
- This PR is huge (71 commits) and summarize several completely parallel developments (new stats model, new modeling capa, performances, testing, etc). Hence, I use a `Merge` instead of `Rebase` to keep track of changes coming from this huge diff.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants