TimeCopilot · AzulGarza · Sep 18, 2026 · Sep 17, 2026 · Sep 17, 2026 · Sep 17, 2026
diff --git a/README.md b/README.md
@@ -84,7 +84,8 @@ Licenses verified against Hugging Face model cards. Check the model card for you
 <details><summary><strong>What this means for production</strong></summary>
 
 - **Apache-2.0**, **MIT**: generally fine for commercial production (retain notices; T0† is also gated on Hugging Face: accept terms and set `HF_TOKEN`).
-- **CC-BY-NC-4.0** (Moirai), **CC-BY-NC-SA-4.0** (PatchTST-FM): **non-commercial** only; not for revenue-generating production without a separate agreement from the rights holder.
+- **CC-BY-NC-4.0** (Moirai), **CC-BY-NC-SA-4.0** (PatchTST-FM r1 checkpoints): **non-commercial** only; not for revenue-generating production without a separate agreement from the rights holder.
+- **Apache-2.0 / OpenMDW 1.0** (Granite PatchTST-FM r2): commercial-friendly; choose either license per the [model card](https://huggingface.co/ibm-granite/granite-timeseries-patchtst-fm-r2).
 - **TimesFM 3.0**: weights for `google/timesfm-3.0-pytorch` are [non-commercial](https://huggingface.co/google/timesfm-3.0-pytorch/blob/main/LICENSE); TimesFM 1.0–2.5 checkpoints remain Apache-2.0.
 - **TabPFN NC**‡: TabPFN-2.6+ and TabPFN-3 weights are non-commercial; production requires a [Prior Labs commercial license or API](https://docs.priorlabs.ai/models). First LOCAL use requires accepting terms at [ux.priorlabs.ai](https://ux.priorlabs.ai) (`TABPFN_TOKEN`). TabPFN-3 is LOCAL-only today.
 - **Community / Apache-2.0** (TiRex): TiRex 1.0 uses the [NXAI Community License](https://huggingface.co/NX-AI/TiRex/blob/main/LICENSE) (commercial limits for large enterprises); TiRex 2.0 is Apache-2.0.
@@ -101,7 +102,7 @@ Some models require specific Python versions (e.g. FlowState 3.11-3.13, TabPFN &
 - **Chronos:** `amazon/chronos-t5-{tiny,mini,small,base,large}`, `amazon/chronos-bolt-{tiny,mini,small,base}`, `amazon/chronos-2`
 - **FlowState:** `ibm-research/flowstate`, `ibm-granite/granite-timeseries-flowstate-r1`
 - **Moirai:** `Salesforce/moirai-{1.0,1.1,2.0}-R-{small,base,large}`, `Salesforce/moirai-moe-1.0-R-*`
-- **PatchTST-FM:** `ibm-research/patchtst-fm-r1`
+- **PatchTST-FM:** `ibm-research/patchtst-fm-r1`, `ibm-granite/granite-timeseries-patchtst-fm-r1`, `ibm-granite/granite-timeseries-patchtst-fm-r2` (r2 requires `granite-tsfm>=0.3.9`)
 - **Sundial:** `thuml/sundial-base-128m`
 - **T0:** `theforecastingcompany/t0-alpha`, `theforecastingcompany/t0-beta` (requires `tfc-t0>=0.5.0`)
 - **TabPFN:** default `tabpfn-v2-regressor-2noar4o2.ckpt`; TabPFN-3 via `model_path="tabpfn-v3-regressor-v3_20260506_timeseries.ckpt"` (LOCAL, `TABPFN_TOKEN`); or CLIENT mode with a Prior Labs API token

diff --git a/docs/changelogs/v0.1.8.md b/docs/changelogs/v0.1.8.md
@@ -1,5 +1,25 @@
 ### Features
 
+* **Granite PatchTST-FM r2 support**: The shared [`PatchTSTFM`](../api/models/foundation/models.md#foundationforecast.models.patchtst_fm.PatchTSTFM) wrapper now supports [`ibm-granite/granite-timeseries-patchtst-fm-r2`](https://huggingface.co/ibm-granite/granite-timeseries-patchtst-fm-r2) (~385M). r2 adds Conformer blocks, overlapping patches (`patch_stride=8`), and a 99-quantile head on top of r1. Requires `granite-tsfm>=0.3.9`. GIFT-Eval CI replicates `Granite-PatchTST-FM-r2` on `m4_weekly/short`.
+
+    ```python
+    import pandas as pd
+    from foundationforecast.models.patchtst_fm import PatchTSTFM
+
+    df = pd.read_csv(
+        "https://timecopilot.s3.amazonaws.com/public/data/air_passengers.csv",
+        parse_dates=["ds"],
+    )
+    model = PatchTSTFM(
+        repo_id="ibm-granite/granite-timeseries-patchtst-fm-r2",
+        alias="Granite-PatchTST-FM-r2",
+        context_length=8192,
+        batch_size=32,
+    )
+    fcst = model.forecast(df, h=12, freq="MS")
+    assert "Granite-PatchTST-FM-r2" in fcst.columns
+    ```
+
 * **T0 beta support**: The shared [`T0`](../api/models/foundation/models.md#foundationforecast.models.t0.T0) wrapper now works with [`theforecastingcompany/t0-beta`](https://huggingface.co/theforecastingcompany/t0-beta) (~256M). Register it in gift-eval `models.yaml`; GIFT-Eval CI still replicates `t0-alpha` until leaderboard results exist for beta.
 
     ```python
@@ -25,6 +45,7 @@
 ### Dependencies
 
 * Bumped `tfc-t0` from `>=0.2.3` to `>=0.5.0` (required for t0-beta; still supports t0-alpha).
+* Replaced `timecopilot-granite-tsfm` with `granite-tsfm>=0.3.9` (required for PatchTST-FM r2 overlapping-patch inference).
 
 ---
 

diff --git a/docs/contributing.md b/docs/contributing.md
@@ -88,10 +88,13 @@ When adding a new foundation model:
 **foundationforecast** uses forked Python packages maintained under custom names on PyPI:
 
 - **chronos-forecasting** — [`timecopilot-chronos-forecasting`](https://pypi.org/project/timecopilot-chronos-forecasting/)
-- **granite-tsfm** — [`timecopilot-granite-tsfm`](https://pypi.org/project/timecopilot-granite-tsfm/)
 - **timesfm** — [`timecopilot-timesfm`](https://pypi.org/project/timecopilot-timesfm/)
 - **tirex** — [`timecopilot-tirex`](https://pypi.org/project/timecopilot-tirex/)
 - **toto** — [`timecopilot-toto`](https://pypi.org/project/timecopilot-toto/)
 - **uni2ts** — [`timecopilot-uni2ts`](https://pypi.org/project/timecopilot-uni2ts/)
 
+FlowState and PatchTST-FM use upstream [`granite-tsfm`](https://pypi.org/project/granite-tsfm/)
+(`>=0.3.9`; PatchTST-FM r2 requires overlapping-patch support). The former
+`timecopilot-granite-tsfm` fork is no longer a dependency.
+
 See the [TimeCopilot contributing guide](https://timecopilot.dev/community/contributing/) for upstream fork links.
diff --git a/docs/model-hub.md b/docs/model-hub.md
@@ -39,7 +39,7 @@ Below is the list of available foundation models. Click a model name for API det
 - [Chronos](api/models/foundation/models.md#foundationforecast.models.chronos.Chronos) ([arXiv:2403.07815](https://arxiv.org/abs/2403.07815))
 - [FlowState](api/models/foundation/models.md#foundationforecast.models.flowstate.FlowState) ([arXiv:2508.05287](https://arxiv.org/abs/2508.05287)) — Python 3.11–3.13
 - [Moirai](api/models/foundation/models.md#foundationforecast.models.moirai.Moirai) ([arXiv:2402.02592](https://arxiv.org/abs/2402.02592))
-- [PatchTST-FM](api/models/foundation/models.md#foundationforecast.models.patchtst_fm.PatchTSTFM) ([arXiv:2602.06909](https://arxiv.org/abs/2602.06909)) — Python 3.11–3.13
+- [PatchTST-FM](api/models/foundation/models.md#foundationforecast.models.patchtst_fm.PatchTSTFM) ([arXiv:2602.06909](https://arxiv.org/abs/2602.06909)) — Python 3.11–3.13; [`ibm-research/patchtst-fm-r1`](https://huggingface.co/ibm-research/patchtst-fm-r1), [`ibm-granite/granite-timeseries-patchtst-fm-r1`](https://huggingface.co/ibm-granite/granite-timeseries-patchtst-fm-r1) (CC-BY-NC-SA-4.0), and [`ibm-granite/granite-timeseries-patchtst-fm-r2`](https://huggingface.co/ibm-granite/granite-timeseries-patchtst-fm-r2) (Apache-2.0 / OpenMDW 1.0; requires `granite-tsfm>=0.3.9`)
 - [Sundial](api/models/foundation/models.md#foundationforecast.models.sundial.Sundial) ([arXiv:2502.00816](https://arxiv.org/pdf/2502.00816))
 - [T0](api/models/foundation/models.md#foundationforecast.models.t0.T0) ([t0-alpha](https://huggingface.co/theforecastingcompany/t0-alpha), [t0-beta](https://huggingface.co/theforecastingcompany/t0-beta); requires `tfc-t0>=0.5.0` for beta) — Python 3.11–3.13
 - [TabPFN](api/models/foundation/models.md#foundationforecast.models.tabpfn.TabPFN) ([arXiv:2501.02945](https://arxiv.org/abs/2501.02945)) — Python 3.10–3.12; TabPFN-2 (default) and TabPFN-3 via `model_path`; TabPFN-3 is LOCAL-only; weights use the [TabPFN NC license](https://ux.priorlabs.ai) (`TABPFN_TOKEN`)

diff --git a/experiments/gift-eval/README.md b/experiments/gift-eval/README.md
@@ -44,9 +44,10 @@ uv run python -m src.runners.run_model \
 
 ## CI subset
 
-[`configs/ci_subset.yaml`](configs/ci_subset.yaml) defines **10 jobs**: Chronos on
+[`configs/ci_subset.yaml`](configs/ci_subset.yaml) defines **13 jobs**: Chronos on
 `m4_weekly/short` and `m4_hourly/short`, plus one representative `model_key` each for
-TimesFM, TiRex, Moirai, Toto, FlowState, PatchTST-FM, T0, and Tafsut (all on `m4_weekly/short`).
+TimesFM 2.5/3.0, TiRex 1.1 and TiRex-2-Zeroshot, Moirai, Toto, FlowState, PatchTST-FM r1,
+Granite PatchTST-FM r2, T0, and Tafsut (all on `m4_weekly/short`).
 Each job runs on Modal GPU and is **HF-verified** in pytest (metrics must match the
 official GIFT-Eval reference CSV).
 

diff --git a/experiments/gift-eval/configs/ci_subset.yaml b/experiments/gift-eval/configs/ci_subset.yaml
@@ -1,5 +1,6 @@
 # Representative CI subset for replication verification (Chronos on two datasets).
 # Every job is HF-verified in tests/test_replication.py after Modal run_ci.
+# New models (TiRex-2, Granite PatchTST-FM r2) are added alongside existing entries.
 jobs:
   - model_key: amazon--chronos-bolt-small
     dataset_name: m4_weekly
@@ -16,6 +17,9 @@ jobs:
   - model_key: NX-AI--TiRex-1.1-gifteval
     dataset_name: m4_weekly
     term: short
+  - model_key: NX-AI--TiRex-2-gifteval-zs
+    dataset_name: m4_weekly
+    term: short
   - model_key: Salesforce--moirai-1.1-R-base
     dataset_name: m4_weekly
     term: short
@@ -28,6 +32,9 @@ jobs:
   - model_key: ibm-research--patchtst-fm-r1
     dataset_name: m4_weekly
     term: short
+  - model_key: ibm-granite--granite-timeseries-patchtst-fm-r2
+    dataset_name: m4_weekly
+    term: short
   - model_key: theforecastingcompany--t0-alpha
     dataset_name: m4_weekly
     term: short

diff --git a/experiments/gift-eval/configs/models.yaml b/experiments/gift-eval/configs/models.yaml
@@ -242,6 +242,15 @@ models:
       repo_id: ibm-granite/granite-timeseries-patchtst-fm-r1
       batch_size: 32
 
+  ibm-granite--granite-timeseries-patchtst-fm-r2:
+    class: foundationforecast.models.patchtst_fm.PatchTSTFM
+    reference_slug: Granite-PatchTST-FM-r2
+    max_length: 8192
+    kwargs:
+      repo_id: ibm-granite/granite-timeseries-patchtst-fm-r2
+      context_length: 8192
+      batch_size: 32
+
   # --- Other foundation models ---
   theforecastingcompany--t0-alpha:
     class: foundationforecast.models.t0.T0

diff --git a/experiments/gift-eval/tests/test_models_config.py b/experiments/gift-eval/tests/test_models_config.py
@@ -8,3 +8,13 @@ def test_t0_beta_registered_in_models_yaml():
     assert entry["reference_slug"] is None
     assert entry["kwargs"]["repo_id"] == "theforecastingcompany/t0-beta"
     assert entry["kwargs"]["alias"] == "t0-beta"
+
+
+def test_granite_patchtst_r2_registered_in_models_yaml():
+    models = load_models_config()
+    entry = models["ibm-granite--granite-timeseries-patchtst-fm-r2"]
+    assert entry["class"] == "foundationforecast.models.patchtst_fm.PatchTSTFM"
+    assert entry["reference_slug"] == "Granite-PatchTST-FM-r2"
+    assert entry["max_length"] == 8192
+    assert entry["kwargs"]["repo_id"] == "ibm-granite/granite-timeseries-patchtst-fm-r2"
+    assert entry["kwargs"]["context_length"] == 8192