AzulGarza · Oct 26, 2026 · Oct 26, 2026
diff --git a/README.md b/README.md
@@ -143,11 +143,11 @@ Some models require specific Python versions (e.g. FlowState 3.11-3.13, TabPFN &
 
 ## News
 
-Recent additions and checkpoints (newest first). See the [changelog](https://github.com/TimeCopilot/foundationforecast/blob/main/docs/changelogs/v0.1.16.md) for full release notes.
+Recent additions and checkpoints (newest first). See the [changelog](https://github.com/TimeCopilot/foundationforecast/blob/main/docs/changelogs/v0.1.17.md) for full release notes.
 
 | When | What | Checkpoint / detail |
 |------|------|---------------------|
-| Oct 2026 | [**Tabby**](https://huggingface.co/paris-noah/Tabby) | `paris-noah/Tabby` — 99 native quantiles; `trust_remote_code`; CC-BY-NC-4.0 |
+| Oct 2026 | [**Tabby**](https://huggingface.co/paris-noah/Tabby) | `paris-noah/Tabby` — 99 native quantiles; `trust_remote_code`; CC-BY-NC-4.0; GIFT-Eval CI ([v0.1.17](https://github.com/TimeCopilot/foundationforecast/blob/main/docs/changelogs/v0.1.17.md)) |
 | Oct 2026 | [**TimeH2O**](https://huggingface.co/h2oai/timeh2o-v1) | `h2oai/timeh2o-v1`, `h2oai/timeh2o-v1-tiny` — requires `h2ofm>=0.43.0` (Python 3.12+); native `X_df`; GIFT-Eval and fev-bench CI ([v0.1.16](https://github.com/TimeCopilot/foundationforecast/blob/main/docs/changelogs/v0.1.16.md)) |
 | Oct 2026 | [**Timestral**](https://huggingface.co/collections/mistralai/timestral) | `mistralai/Timestral-{5M,28M,123M,209M}` — requires `timecopilot-timestral>=0.1.0`; native `X_df`; GIFT-Eval and fev-bench CI ([v0.1.15](https://github.com/TimeCopilot/foundationforecast/blob/main/docs/changelogs/v0.1.15.md)) |
 | Oct 2026 | [**T0 gamma**](https://huggingface.co/theforecastingcompany/t0-gamma) | `theforecastingcompany/t0-gamma` — requires `tfc-t0[torch]>=0.6.0`; GIFT-Eval and fev-bench CI ([v0.1.14](https://github.com/TimeCopilot/foundationforecast/blob/main/docs/changelogs/v0.1.14.md)) |

diff --git a/docs/changelogs/index.md b/docs/changelogs/index.md
@@ -2,6 +2,7 @@
 
 Welcome to the FoundationForecast Changelog. Here you will find a list of changes, updates, and improvements for each release. Explore the details of each version below.
 
+- [v0.1.17](v0.1.17.md)
 - [v0.1.16](v0.1.16.md)
 - [v0.1.15](v0.1.15.md)
 - [v0.1.14](v0.1.14.md)

diff --git a/docs/changelogs/v0.1.17.md b/docs/changelogs/v0.1.17.md
@@ -0,0 +1,17 @@
+### Features
+
+* **Tabby**: The [`Tabby`](https://docs.timecopilot.dev/foundationforecast/api/models/foundation/models/#foundationforecast.models.tabby.Tabby) wrapper supports Huawei Paris Noah's Ark Lab's [`paris-noah/Tabby`](https://huggingface.co/paris-noah/Tabby) ([arXiv:2609.13956](https://arxiv.org/abs/2609.13956)), a ~147M prompted patch transformer that emits 99 native quantiles (0.01–0.99) from up to 8,192 observations. The wrapper follows the published inference path: 8,096-step context, `context_mask` and `past_is_pad` passed to the forward, and a `precision` option (`"bf16"` autocast on CUDA by default, `"fp32"` to disable it). The checkpoint loads with `trust_remote_code` and is released under CC BY-NC 4.0. See [#76](https://github.com/TimeCopilot/foundationforecast/pull/76).
+
+### Tests
+
+* **Tabby**: Mocked tests for quantile mapping, context truncation and precision handling, a live test on `paris-noah/Tabby`, and cross-model contract coverage through `tests/models/conftest.py`. See [#76](https://github.com/TimeCopilot/foundationforecast/pull/76).
+* **GIFT-Eval replication**: `paris-noah--Tabby` is registered against the official `results/Tabby` entry (MASE 0.688, CRPS 0.4758), with the `m4_weekly/short` subset added to CI. The leaderboard flags that entry `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. There is no fev-bench entry, since Tabby is univariate and has no official result there. See [#76](https://github.com/TimeCopilot/foundationforecast/pull/76).
+
+### Documentation
+
+* **Tabby family notebook**: [`tabby-family` example notebook](../examples/tabby-family.ipynb) cross-validates Tabby with its published settings on event pageviews, with prediction intervals and executed outputs. See [#76](https://github.com/TimeCopilot/foundationforecast/pull/76).
+* **Model hub and README**: Document the Tabby checkpoint, `precision` option, CC BY-NC 4.0 license and `trust_remote_code` requirement; add Tabby to the supported-models table with the Huawei Noah's Ark Lab logo.
+
+---
+
+**Full Changelog**: https://github.com/TimeCopilot/foundationforecast/compare/v0.1.16...v0.1.17
diff --git a/experiments/fev-bench/uv.lock b/experiments/fev-bench/uv.lock
diff --git a/experiments/gift-eval/uv.lock b/experiments/gift-eval/uv.lock
diff --git a/mkdocs.yml b/mkdocs.yml
@@ -51,6 +51,7 @@ nav:
       - Contributing: contributing.md
   - Changelogs:
       - changelogs/index.md
+      - changelogs/v0.1.17.md
       - changelogs/v0.1.16.md
       - changelogs/v0.1.15.md
       - changelogs/v0.1.14.md

diff --git a/pyproject.toml b/pyproject.toml
@@ -75,7 +75,7 @@ license = "Apache-2.0"
 name = "foundationforecast"
 readme = "README.md"
 requires-python = ">=3.10"
-version = "0.1.16"
+version = "0.1.17"
 
 [project.optional-dependencies]
 plot = [

diff --git a/uv.lock b/uv.lock