---
pipeline_tag: image-classification
license: apache-2.0
base_model: timm/pit_b_distilled_224.in1k
library_name: zeromodels
tags:
- keras
- zeromodels
- image-classification
- pit
- backbone
- arxiv:2103.16427
- pytorch
- jax
- tf
---

## ***See [our collection](https://huggingface.co/collections/zeromodels/pit-6a8eaec91c9608b135a564b6) for all versions of PiT.***

# Run PiT with Keras 3: JAX, PyTorch, or TensorFlow

[![GitHub](https://img.shields.io/badge/GitHub-ZeroModels-black?logo=github)](https://github.com/IMvision12/ZeroModels) [![Docs](https://img.shields.io/badge/Docs-PiT-blue)](https://imvision12.github.io/ZeroModels/classification_backbones/) [![Collection](https://img.shields.io/badge/HF-PiT%20collection-yellow)](https://huggingface.co/collections/zeromodels/pit-6a8eaec91c9608b135a564b6)

# zeromodels/pit_b_distilled_224_in1k

Paper: [Rethinking Spatial Dimensions of Vision Transformers (arXiv:2103.16427)](https://arxiv.org/abs/2103.16427) · [HF Papers](https://huggingface.co/papers/2103.16427)

Pooling-based Vision Transformer (PiT) reshapes spatial dimensions across stages like a CNN. Classifier or multi-stage backbone.

For more details on the model, please go to the upstream [model card](https://huggingface.co/timm/pit_b_distilled_224.in1k).

Pure-**Keras 3** conversion of [`timm/pit_b_distilled_224.in1k`](https://huggingface.co/timm/pit_b_distilled_224.in1k) for [zeromodels](https://github.com/IMvision12/ZeroModels). One implementation runs unmodified on **TensorFlow / Torch / JAX**.

This is an **image-classification / backbone** checkpoint (`PiTImageClassify` / `PiTModel`).

## ✨ Quick start

```python
import os

os.environ["KERAS_BACKEND"] = "torch"  # or "jax" / "tensorflow"

from PIL import Image
from zeromodels.models.pit import PiTImageClassify, PiTModel, PiTImageProcessor

model = PiTImageClassify.from_weights("zeromodels/pit_b_distilled_224_in1k")
processor = PiTImageProcessor.from_weights("zeromodels/pit_b_distilled_224_in1k")

image = Image.open("your_image.jpg").convert("RGB")
pixels = processor(image)  # resize + normalize (normalization lives in the processor)
logits = model(pixels, training=False)
print(logits.shape)  # (1, num_classes)

# Feature extraction: the backbone without the classifier head
backbone = PiTModel.from_weights("zeromodels/pit_b_distilled_224_in1k", as_backbone=True)
features = backbone(pixels, training=False)
```

Load any PiT variant the same way with `from_weights("zeromodels/<variant>")`:

| Variant | Hub |
|---|---|
| `pit_b_224_in1k` | [`zeromodels/pit_b_224_in1k`](https://huggingface.co/zeromodels/pit_b_224_in1k) |
| `pit_b_distilled_224_in1k` | [`zeromodels/pit_b_distilled_224_in1k`](https://huggingface.co/zeromodels/pit_b_distilled_224_in1k) |
| `pit_s_224_in1k` | [`zeromodels/pit_s_224_in1k`](https://huggingface.co/zeromodels/pit_s_224_in1k) |
| `pit_s_distilled_224_in1k` | [`zeromodels/pit_s_distilled_224_in1k`](https://huggingface.co/zeromodels/pit_s_distilled_224_in1k) |
| `pit_ti_224_in1k` | [`zeromodels/pit_ti_224_in1k`](https://huggingface.co/zeromodels/pit_ti_224_in1k) |
| `pit_ti_distilled_224_in1k` | [`zeromodels/pit_ti_distilled_224_in1k`](https://huggingface.co/zeromodels/pit_ti_distilled_224_in1k) |
| `pit_xs_224_in1k` | [`zeromodels/pit_xs_224_in1k`](https://huggingface.co/zeromodels/pit_xs_224_in1k) |
| `pit_xs_distilled_224_in1k` | [`zeromodels/pit_xs_distilled_224_in1k`](https://huggingface.co/zeromodels/pit_xs_distilled_224_in1k) |

## Tips

- Set `KERAS_BACKEND` **before** importing Keras / zeromodels.
- `PiTImageClassify` returns class logits; `PiTModel` returns features (`as_backbone=True` for multi-scale stages).
- See [docs](https://imvision12.github.io/ZeroModels/classification_backbones/) and [Loading Weights](https://imvision12.github.io/ZeroModels/loading_weights/).
- Upstream / timm checkpoints: `PiTImageClassify.from_weights("hf:timm/pit_b_distilled_224.in1k")`.

## Special Thanks

A huge thank you to the PiT authors and the timm / Hub communities for creating and releasing these models.

License: see YAML `license` (usually matches the upstream checkpoint).