Add v3 OOD checkpoints + README

README.md

@@ -65,6 +65,8 @@ The following specialized checkpoints are available:
 | [`tabpfn-v3-classifier-v3_20260417_multiclass.ckpt`](https://huggingface.co/Prior-Labs/tabpfn_3/blob/main/tabpfn-v3-classifier-v3_20260417_multiclass.ckpt) | Classification | Specialized for multiclass classification for datasets with <200k rows |
 | [`tabpfn-v3-regressor-v3_20260417_mediumdata.ckpt`](https://huggingface.co/Prior-Labs/tabpfn_3/blob/main/tabpfn-v3-regressor-v3_20260417_mediumdata.ckpt) | Regression | Specialized for regression for datasets with <100k rows and with alternative preprocessing |
 | [`tabpfn-v3-regressor-v3_20260506_timeseries.ckpt`](https://huggingface.co/Prior-Labs/tabpfn_3/blob/main/tabpfn-v3-regressor-v3_20260506_timeseries.ckpt) | Regression / Time-series forecasting | Fine-tuned on synthetic time-series data; used by default in TabPFN-TS-3 |
+| [`tabpfn-v3-classifier-v3_ood.ckpt`](https://huggingface.co/Prior-Labs/tabpfn_3/blob/main/tabpfn-v3-classifier-v3_ood.ckpt) | Classification | Same weights as the default classifier but with OOD-robust preprocessors bundled (`squashing_scaler_max10` + `none`). Useful when test inputs may fall outside the training distribution. |
+| [`tabpfn-v3-regressor-v3_ood.ckpt`](https://huggingface.co/Prior-Labs/tabpfn_3/blob/main/tabpfn-v3-regressor-v3_ood.ckpt) | Regression | Same weights as the default regressor but with OOD-robust preprocessors bundled (`quantile_uni_extrapolate` + `squashing_scaler_max10`). Linearly extrapolates past the training range instead of clamping. Requires `tabpfn` from the public main branch (post-#971 merge). |
 
 To use one of these checkpoints, pass its filename via `model_path`:
 

tabpfn-v3-classifier-v3_ood.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f973d994ece09f9e28e9319d644bcff3cb0bec9c3170e67e14e71f53df92d18b
+size 212815699

tabpfn-v3-regressor-v3_ood.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fe1a5aa90cdd7f06e8736a68015b1bb570a413bd302de2d1038364272a8c24d6
+size 233300137