Skip to content

[DRAFT] Mh/latent rmse diagnostics on develop ssl v1 - #2703

Draft
moritzhauschulz wants to merge 16 commits into
ecmwf:develop-ssl-diffusion-v1from
moritzhauschulz:mh/latent-rmse-diagnostics-on-develop-ssl-v1
Draft

[DRAFT] Mh/latent rmse diagnostics on develop ssl v1#2703
moritzhauschulz wants to merge 16 commits into
ecmwf:develop-ssl-diffusion-v1from
moritzhauschulz:mh/latent-rmse-diagnostics-on-develop-ssl-v1

Conversation

@moritzhauschulz

Copy link
Copy Markdown
Contributor

Description

[DRAFT]

Attempt at combining https://github.com/ecmwf/WeatherGenerator/pull/2702/changes and https://github.com/ecmwf/WeatherGenerator/pull/2683/changes and merging with SSL branch. Untested.

Issue Number

Is this PR a draft? YES!

Checklist before asking for review

  • I have performed a self-review of my code
  • My changes comply with basic sanity checks:
    • I have fixed formatting issues with ./scripts/actions.sh lint
    • I have run unit tests with ./scripts/actions.sh unit-test
    • I have documented my code and I have updated the docstrings.
    • I have added unit tests, if relevant
  • I have tried my changes with data and code:
    • I have run the integration tests with ./scripts/actions.sh integration-test
    • (bigger changes) I have run a full training and I have written in the comment the run_id(s): launch-slurm.py --time 60
    • (bigger changes and experiments) I have shared a hegdedoc in the github issue with all the configurations and runs for this experiments
  • I have informed and aligned with people impacted by my change:
    • for config changes: the MatterMost channels and/or a design doc
    • for changes of dependencies: the MatterMost software development channel

@moritzhauschulz
moritzhauschulz changed the base branch from develop to develop-ssl-diffusion-v1 July 30, 2026 16:47
@github-actions github-actions Bot added eval anything related to the model evaluation pipeline model Related to model training or definition (not generic infra) labels Jul 30, 2026
@moritzhauschulz
moritzhauschulz force-pushed the mh/latent-rmse-diagnostics-on-develop-ssl-v1 branch 2 times, most recently from 5648594 to 457a3df Compare August 11, 2026 13:00
moritzhauschulz and others added 13 commits August 14, 2026 09:52
Squash of the mh/...-latent-rmse-and-diagnostics-ssl feature commits
(1785ef3..4356ee6) resolved onto develop-ssl-diffusion-v1. Adds:
- Latent-rollout RMSE diagnostic: record_latent_rollout side channel +
  isolated batch.latent_rmse_source truth encode (utils/latent_rmse.py).
- Per-ODE-step map & spectrum diagnostics (inference_diagnostics.py,
  inference_spectra.py, evaluate/scores/psd.py; diag_* config keys).
- Harmonized diffusion sampling diagnostics (per-element rmse panel, 4-panel).

Conflict resolutions vs develop-ssl: encoder 3-value return threaded through the
chunked truth-latent encode; develop's LATENT_CONDITIONING_TOKENS key kept;
_get_data_windows 3-return reconciled with teacher_time_offset windows;
fe_diffusion_num_inference_steps config key; MapPanelPlotter adapted to develop's
eval Plotter API (norm built locally). 28 diagnostics/spectra tests pass.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
@moritzhauschulz
moritzhauschulz force-pushed the mh/latent-rmse-diagnostics-on-develop-ssl-v1 branch from 457a3df to 8a9587e Compare August 17, 2026 15:26
@github-actions github-actions Bot added data Anything related to the datasets used in the project infra Issues related to infrastructure labels Aug 17, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

data Anything related to the datasets used in the project eval anything related to the model evaluation pipeline infra Issues related to infrastructure model Related to model training or definition (not generic infra)

Projects

Status: No status

Development

Successfully merging this pull request may close these issues.

2 participants