[AzulGarza]

https://entire.com/gh/timecopilot/-/trails/1/changes/foundationforecast/1

Summary

• Add a Tabby wrapper for paris-noah/Tabby (AutoModel + trust_remote_code), 99 native quantiles, no native exog.
• Cross-model contract via tests/models/conftest.py, plus focused forecast/level tests.
• Docs: model hub, API page, README matrix (CC-BY-NC-4.0), exog table, and tabby-family notebook.

GIFT-Eval / fev-bench CI is not in this PR. Both public Tabby submissions are marked testdata_leakage: Yes.

Test plan

• uv run pre-commit run --all-files
• pytest tests/models/test_tabby.py (mocked + live paris-noah/Tabby)
• CI test / lint / test-docs / build-docs on the draft

Made with Cursor

GIFT-Eval replication

• paris-noah--Tabby in experiments/gift-eval/configs/models.yaml, verified against the official results/Tabby entry. That entry is the current paris-noah/Tabby release (seed-42 prompt, 128-step minimum forecast span): its CSV gives MASE 0.688 / CRPS 0.4758, matching the model card. The leaderboard flags it testdata_leakage: Yes because the prompt was tuned on GIFT-Eval training tasks; the harness verifies the wrapper reproduces the official numbers regardless of that flag.
• Settings match the official evaluation script: 8,096-step context, bf16 autocast, context_mask and past_is_pad passed to the forward. CI subset: m4_weekly/short.
• No fev-bench entry: there is no official Tabby result there, and the fev-bench harness only runs covariate tasks (Tabby is univariate).

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.

An organization owner can choose the repository organization as the billing source for reviews requested by members with a paid Copilot license. Learn about Copilot code review billing.

[codspeed]

Merging this PR will not alter performance

1 note

ℹ️ 13 benchmarks have no flamegraph

The flamegraph wasn't generated: the symbols of the profiling trace could not be resolved, because no information about the language frames was collected.

See affected benchmarks

✅ 17 untouched benchmarks

---

_{Comparing feat/tabby (adb7563) with main (5d833ee)}

[codecov]

Codecov Report

❌ Patch coverage is 98.71795% with 1 line in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
foundationforecast/models/tabby.py	98.70%	1 Missing ⚠️

📢 Thoughts on this report? Let us know!

[codspeed]

Unable to generate the performance report

Your organization has exceeded its Macro Runners budget. Because this run uses tokenless mode, CodSpeed needs to wait for the GitHub Actions workflow to complete to validate the upload. Since the walltime jobs will never run, the workflow stays pending and we are unable to process this run.

To run these benchmarks again, update your plan or increase your Macro Runners budget.

If you need help, please contact us at support@codspeed.io or on Discord.