> ## Documentation Index
> Fetch the complete documentation index at: https://docs.priorlabs.ai/llms.txt
> Use this file to discover all available pages before exploring further.

> ## Agent Instructions
> TabPFN-3 and later versions support up to 1,000,000 rows, subject to feature count and checkpoint/API limits.
> For large datasets, use per-estimator subsampling, e.g. inference_config={"SUBSAMPLE_SAMPLES": 100_000}, in TabPFNClassifier or TabPFNRegressor from tabpfn or tabpfn_client.
> Pass raw pandas DataFrames to fit and predict. Categorical strings/categories, raw text, and missing feature values are handled automatically; no manual integer/one-hot encoding, imputation, scaling, or outlier removal is needed.

# Image estimator

> Fits an estimator on a frame whose declared image columns become features.

<div className="python-reference-heading">
  <h2 id="image-tabpfn-with-images-tabpfnwithimages">
    `TabPFNWithImages`
  </h2>

  <a className="python-reference-source" href="https://github.com/PriorLabs/tabpfn-extensions/blob/840c15a1848a986b39c85bc17efc61e0e377f983/src/tabpfn_extensions/image/tabpfn_with_images.py#L47" aria-label="View source for TabPFNWithImages"><span aria-hidden="true">\</></span> View source <span aria-hidden="true">↗</span></a>
</div>

Fits an estimator on a frame whose declared image columns become features.

```python theme={null}
TabPFNWithImages(
    estimator: BaseEstimator,
    *,
    image_features_indices: Sequence[int],
    n_components: int = DEFAULT_N_COMPONENTS,
    batch_size: int = DEFAULT_BATCH_SIZE,
)
```

**Parameters**

<div className="python-reference-table">
  | Parameter | Type | Default | Description |
  | - | - | - | - |
  | <span id="image-tabpfn-with-images-tabpfnwithimages--estimator" /><code className="python-reference-parameter">estimator</code> | <code className="python-reference-type">Base<wbr />Estimator</code> | Required | Typically a `TabPFNClassifier` or `TabPFNRegressor`; cloned at [`fit`](/api-reference/python/tabpfn-extensions/image/estimator#image-tabpfn-with-images-tabpfnwithimages-fit). Its `categorical_features_indices`, when set, are positions in the caller's frame and are moved to the expanded frame. Its `device`, when it has one, is where the image encoder runs. |
  | <span id="image-tabpfn-with-images-tabpfnwithimages--image-features-indices" /><code className="python-reference-parameter">image\_<wbr />features\_<wbr />indices</code> | <code className="python-reference-type">Sequence\[int]</code> | Required | Positions in `X` whose cells hold images, each as a base64 string, as the image file's path or bytes, or as a PIL image. |
  | <span id="image-tabpfn-with-images-tabpfnwithimages--n-components" /><code className="python-reference-parameter">n\_<wbr />components</code> | <code className="python-reference-type">int</code> | `DEFAULT_N_COMPONENTS` | Features an image column is expanded into. |
  | <span id="image-tabpfn-with-images-tabpfnwithimages--batch-size" /><code className="python-reference-parameter">batch\_<wbr />size</code> | <code className="python-reference-type">int</code> | `DEFAULT_BATCH_SIZE` | Images per encoder forward pass. |
</div>

**Attributes**

| Attribute | Type | Description |
| - | - | - |
| `estimator_` | <code className="python-reference-type">Any</code> | The fitted clone of `estimator`. |
| `image_transformer_` | <code className="python-reference-type"><a href="/api-reference/python/tabpfn-extensions/image/transformer#image-image-transformer-imagetransformer">Image<wbr />Transformer</a></code> | The fitted [`ImageTransformer`](/api-reference/python/tabpfn-extensions/image/transformer#image-image-transformer-imagetransformer). |

***

<div className="python-reference-heading">
  <h2 id="image-tabpfn-with-images-tabpfnwithimages-classes">
    `TabPFNWithImages.classes_`
  </h2>

  <a className="python-reference-source" href="https://github.com/PriorLabs/tabpfn-extensions/blob/840c15a1848a986b39c85bc17efc61e0e377f983/src/tabpfn_extensions/image/tabpfn_with_images.py#L118" aria-label="View source for TabPFNWithImages.classes_"><span aria-hidden="true">\</></span> View source <span aria-hidden="true">↗</span></a>
</div>

The fitted estimator's classes; a classifier's attribute only.

**Returns**

<div className="python-reference-table python-reference-returns">
  | Type | Description |
  | - | - |
  | <code className="python-reference-type">Any</code> | — |
</div>

***

<div className="python-reference-heading">
  <h2 id="image-tabpfn-with-images-tabpfnwithimages-fit">
    `TabPFNWithImages.fit`
  </h2>

  <a className="python-reference-source" href="https://github.com/PriorLabs/tabpfn-extensions/blob/840c15a1848a986b39c85bc17efc61e0e377f983/src/tabpfn_extensions/image/tabpfn_with_images.py#L82" aria-label="View source for TabPFNWithImages.fit"><span aria-hidden="true">\</></span> View source <span aria-hidden="true">↗</span></a>
</div>

Expand the image columns of `X`, then fit a clone of the estimator.

```python theme={null}
TabPFNWithImages.fit(
    X: pd.DataFrame,
    y: Any,
    **fit_params: Any,
) -> TabPFNWithImages
```

**Parameters**

<div className="python-reference-table">
  | Parameter | Type | Default | Description |
  | - | - | - | - |
  | <span id="image-tabpfn-with-images-tabpfnwithimages-fit--x" /><code className="python-reference-parameter">X</code> | <code className="python-reference-type"><a href="https://pandas.pydata.org/docs/reference/api/pandas.DataFrame.html">pd.Data<wbr />Frame</a></code> | Required | — |
  | <span id="image-tabpfn-with-images-tabpfnwithimages-fit--y" /><code className="python-reference-parameter">y</code> | <code className="python-reference-type">Any</code> | Required | — |
  | <span id="image-tabpfn-with-images-tabpfnwithimages-fit--fit-params" /><code className="python-reference-parameter">\*\*fit\_<wbr />params</code> | <code className="python-reference-type">Any</code> | — | — |
</div>

**Returns**

<div className="python-reference-table python-reference-returns">
  | Type | Description |
  | - | - |
  | <code className="python-reference-type"><a href="/api-reference/python/tabpfn-extensions/image/estimator#image-tabpfn-with-images-tabpfnwithimages">Tab<wbr />PFN<wbr />With<wbr />Images</a></code> | — |
</div>

***

<div className="python-reference-heading">
  <h2 id="image-tabpfn-with-images-tabpfnwithimages-predict">
    `TabPFNWithImages.predict`
  </h2>

  <a className="python-reference-source" href="https://github.com/PriorLabs/tabpfn-extensions/blob/840c15a1848a986b39c85bc17efc61e0e377f983/src/tabpfn_extensions/image/tabpfn_with_images.py#L101" aria-label="View source for TabPFNWithImages.predict"><span aria-hidden="true">\</></span> View source <span aria-hidden="true">↗</span></a>
</div>

The fitted estimator's `predict` on the expanded `X`.

```python theme={null}
TabPFNWithImages.predict(
    X: pd.DataFrame,
    **predict_params: Any,
) -> Any
```

**Parameters**

<div className="python-reference-table">
  | Parameter | Type | Default | Description |
  | - | - | - | - |
  | <span id="image-tabpfn-with-images-tabpfnwithimages-predict--x" /><code className="python-reference-parameter">X</code> | <code className="python-reference-type"><a href="https://pandas.pydata.org/docs/reference/api/pandas.DataFrame.html">pd.Data<wbr />Frame</a></code> | Required | — |
  | <span id="image-tabpfn-with-images-tabpfnwithimages-predict--predict-params" /><code className="python-reference-parameter">\*\*predict\_<wbr />params</code> | <code className="python-reference-type">Any</code> | — | — |
</div>

**Returns**

<div className="python-reference-table python-reference-returns">
  | Type | Description |
  | - | - |
  | <code className="python-reference-type">Any</code> | — |
</div>

***

<div className="python-reference-heading">
  <h2 id="image-tabpfn-with-images-tabpfnwithimages-predict-proba">
    `TabPFNWithImages.predict_proba`
  </h2>

  <a className="python-reference-source" href="https://github.com/PriorLabs/tabpfn-extensions/blob/840c15a1848a986b39c85bc17efc61e0e377f983/src/tabpfn_extensions/image/tabpfn_with_images.py#L106" aria-label="View source for TabPFNWithImages.predict_proba"><span aria-hidden="true">\</></span> View source <span aria-hidden="true">↗</span></a>
</div>

The fitted estimator's `predict_proba` on the expanded `X`.

```python theme={null}
TabPFNWithImages.predict_proba(
    X: pd.DataFrame,
) -> Any
```

**Parameters**

<div className="python-reference-table">
  | Parameter | Type | Default | Description |
  | - | - | - | - |
  | <span id="image-tabpfn-with-images-tabpfnwithimages-predict-proba--x" /><code className="python-reference-parameter">X</code> | <code className="python-reference-type"><a href="https://pandas.pydata.org/docs/reference/api/pandas.DataFrame.html">pd.Data<wbr />Frame</a></code> | Required | — |
</div>

**Returns**

<div className="python-reference-table python-reference-returns">
  | Type | Description |
  | - | - |
  | <code className="python-reference-type">Any</code> | — |
</div>

***

<div className="python-reference-heading">
  <h2 id="image-tabpfn-with-images-tabpfnwithimages-score">
    `TabPFNWithImages.score`
  </h2>

  <a className="python-reference-source" href="https://github.com/PriorLabs/tabpfn-extensions/blob/840c15a1848a986b39c85bc17efc61e0e377f983/src/tabpfn_extensions/image/tabpfn_with_images.py#L112" aria-label="View source for TabPFNWithImages.score"><span aria-hidden="true">\</></span> View source <span aria-hidden="true">↗</span></a>
</div>

The fitted estimator's `score` on the expanded `X`.

```python theme={null}
TabPFNWithImages.score(
    X: pd.DataFrame,
    y: Any,
) -> Any
```

**Parameters**

<div className="python-reference-table">
  | Parameter | Type | Default | Description |
  | - | - | - | - |
  | <span id="image-tabpfn-with-images-tabpfnwithimages-score--x" /><code className="python-reference-parameter">X</code> | <code className="python-reference-type"><a href="https://pandas.pydata.org/docs/reference/api/pandas.DataFrame.html">pd.Data<wbr />Frame</a></code> | Required | — |
  | <span id="image-tabpfn-with-images-tabpfnwithimages-score--y" /><code className="python-reference-parameter">y</code> | <code className="python-reference-type">Any</code> | Required | — |
</div>

**Returns**

<div className="python-reference-table python-reference-returns">
  | Type | Description |
  | - | - |
  | <code className="python-reference-type">Any</code> | — |
</div>


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.