Languages
View on GitHub →Python95.3%
Jupyter Notebook4.2%
JavaScript0.3%
Makefile0.1%
Batchfile0.1%
CSS0.1%
Commit activity
Last 12 weeks · 54 commits
Community health
4 of 6 standards met
75
✓README✓License✓Contributing✓Code of Conduct○Issue Template○PR Template
Recent PRs & issues
Active · Last activity 5d ago
See all on GitHub →
Summary: Teaches both storage backends to persist and restore the new field on . Second diff in the step_size unification stack (builds on the native support added in D107274057). The underlying SQA column () and the corresponding EntAE schema were added and landed separately (D107108021, D107112306), so this diff is pure encoder/decoder logic — it does not touch the schema. SQA: the encoder writes to the column (and continues writing the legacy column for now; is dropped in a later diff). The decoder prefers and falls back to the legacy column for rows written by older code, passing only one to the constructor (which rejects both being set; it converts to internally). JSON: emits alongside . There is no custom JSON decoder for — the registry maps it directly to the class, and splats the serialized dict straight into the constructor. So flows through automatically, and legacy blobs carrying only still decode via the constructor's back-compat handling. No decoder change is required for JSON. Differential Revision: D107282941
Differential Revision: D61730641
Summary: Refactors the learned imputation argument dispatcher to delegate feature index computation to the dataset's built-in mapping utility. This eliminates duplicated feature-ordering logic and ensures consistency with the canonical ordering convention. Differential Revision: D102197138
Recent fixes
Summary: Prevent from failing model initialization when distinct progression values span less than the minimum width supported by . Preserve the upper bound and expand the lower bound downward to the next representable valid value. Share the minimum-width constant with and cover inferred and configured bounds at and below the threshold. Reviewed By: rpanchal1996 Differential Revision: D116546189
Summary: This diff was automatically generated by the Pyre per-target upgrade tool. It removes or comments that are no longer needed because the underlying type errors have been resolved. Note that it will also aim to ensure type checking runs cleanly, and will add suppressions to existing type errors. #pyreupgrade Differential Revision: D116358917
Summary: Pass to the single-task fallback so shared acquisition behavior such as BONSAI pruning applies regardless of model selection. Document that the kwargs must be compatible with both generators and add constructor coverage. Differential Revision: D116468995
Structured data for AI agents
Repository: facebook/Ax. Description: Adaptive Experimentation Platform Stars: 2791, Forks: 378. Primary language: Python. Languages: Python (95.3%), Jupyter Notebook (4.2%), JavaScript (0.3%), Makefile (0.1%), Batchfile (0.1%). License: MIT. Homepage: https://ax.dev Latest release: 1.3.1 (2mo ago). Open PRs: 100, open issues: 86. Last activity: 5d ago. Community health: 75%. Top contributors: saitcakmak, mpolson64, lena-kashtelyan, esantorella, sdaulton, bernardbeckerman, mgarrard, ldworkin, Balandat, 2timesjay and others.