# BEiT

**Category:** ViT · **Code:**
[`timm/models/beit.py`](https://github.com/huggingface/pytorch-image-models/blob/main/timm/models/beit.py) ·
**Architectures:** 7 · **Pretrained weights:** 13 (9 with ImageNet-1k results)

BEiT and BEiT-v2 Vision Transformers, pretrained with masked image modeling. BEiT predicts the discrete visual tokens of
masked patches, BEiT-v2 uses a vector-quantized visual tokenizer for the targets. The weights here are fine-tuned
classification weights.

## Papers

* [BEiT: BERT Pre-Training of Image Transformers](https://arxiv.org/abs/2106.08254)
* [BEiT v2: Masked Image Modeling with Vector-Quantized Visual Tokenizers](https://arxiv.org/abs/2208.06366)

## Usage

See the [Quickstart](../quickstart) for loading pretrained weights, image preprocessing and fine-tuning, and [Feature
Extraction](../feature_extraction) for using the models as backbones.

```py
>>> import timm
>>> model = timm.create_model('beit_base_patch16_224.in22k_ft_in22k_in1k', pretrained=True)
>>> data_config = timm.data.resolve_model_data_config(model)
```

## Model API support

| Feature | `Beit` |
|---|---|
| [`features_only=True` feature maps](../model_api#intermediate-features) | ✓ (strides 16, 16, 16) |
| [`forward_intermediates()`](../model_api#intermediate-features) | ✓ |
| [`prune_intermediate_layers()`](../model_api#intermediate-features) | ✓ |
| [Any input size, no changes needed](../model_api#input-size) | ✗ |
| [`set_input_size()`](../model_api#input-size) | ✓ |
| [`dynamic_img_size=True`](../model_api#input-size) | ✗ |
| [`set_grad_checkpointing()`](../model_api#training-hooks) | ✓ |
| [Layer-wise LR decay (`group_matcher()`)](../model_api#training-hooks) | ✓ |
| [`reset_classifier()`](../model_api#classifier) | ✓ |

✓ supported, ✓ (square) works for square sizes only, ✗ not supported, – not needed (the model already accepts any input
size). Checked by running each feature on a randomly initialized model (`Beit` with `beit_base_patch16_224` at 224),
other architectures of the same class are expected to behave the same. See the [Model API](../model_api) page for
details.

## Architectures

| Architecture | Params (M) | GMACs | Input size | Pretrained weights |
|---|---:|---:|---:|---:|
| `beit_base_patch16_224` | 86.53 | 17.58 | 224 | 2 |
| `beit_base_patch16_384` | 86.74 | 55.54 | 384 | 1 |
| `beit_large_patch16_224` | 304.43 | 61.60 | 224 | 2 |
| `beit_large_patch16_384` | 305.00 | 191.21 | 384 | 1 |
| `beit_large_patch16_512` | 305.67 | 362.24 | 512 | 1 |
| `beitv2_base_patch16_224` | 86.53 | 17.58 | 224 | 3 |
| `beitv2_large_patch16_224` | 304.43 | 61.60 | 224 | 3 |

GMACs and parameter counts are for the default input size and classifier of each architecture, from the [benchmark
results](https://github.com/huggingface/pytorch-image-models/tree/main/results).

## Pretrained weights

Top-1 is ImageNet-1k validation accuracy from
[`results-imagenet.csv`](https://github.com/huggingface/pytorch-image-models/blob/main/results/results-imagenet.csv) at
the listed eval size, - if the weight has no ImageNet-1k result (e.g. pretraining or image-text weights). Classes is the
size of the pretrained classifier, none for weights without one.

| Weight | Input size | Top-1 (%) | Eval size | Classes | License | Hub |
|---|---:|---:|---:|---:|---|---|
| `beit_base_patch16_224.in22k_ft_in22k` | 224 | - | - | 21841 | apache-2.0 | [timm](https://huggingface.co/timm/beit_base_patch16_224.in22k_ft_in22k) |
| `beit_base_patch16_224.in22k_ft_in22k_in1k` | 224 | 85.238 | 224 | 1000 | apache-2.0 | [timm](https://huggingface.co/timm/beit_base_patch16_224.in22k_ft_in22k_in1k) |
| `beit_base_patch16_384.in22k_ft_in22k_in1k` | 384 | 86.814 | 384 | 1000 | apache-2.0 | [timm](https://huggingface.co/timm/beit_base_patch16_384.in22k_ft_in22k_in1k) |
| `beit_large_patch16_224.in22k_ft_in22k` | 224 | - | - | 21841 | apache-2.0 | [timm](https://huggingface.co/timm/beit_large_patch16_224.in22k_ft_in22k) |
| `beit_large_patch16_224.in22k_ft_in22k_in1k` | 224 | 87.486 | 224 | 1000 | apache-2.0 | [timm](https://huggingface.co/timm/beit_large_patch16_224.in22k_ft_in22k_in1k) |
| `beit_large_patch16_384.in22k_ft_in22k_in1k` | 384 | 88.380 | 384 | 1000 | apache-2.0 | [timm](https://huggingface.co/timm/beit_large_patch16_384.in22k_ft_in22k_in1k) |
| `beit_large_patch16_512.in22k_ft_in22k_in1k` | 512 | 88.576 | 512 | 1000 | apache-2.0 | [timm](https://huggingface.co/timm/beit_large_patch16_512.in22k_ft_in22k_in1k) |
| `beitv2_base_patch16_224.in1k_ft_in1k` | 224 | 85.596 | 224 | 1000 | apache-2.0 | [timm](https://huggingface.co/timm/beitv2_base_patch16_224.in1k_ft_in1k) |
| `beitv2_base_patch16_224.in1k_ft_in22k` | 224 | - | - | 21841 | apache-2.0 | [timm](https://huggingface.co/timm/beitv2_base_patch16_224.in1k_ft_in22k) |
| `beitv2_base_patch16_224.in1k_ft_in22k_in1k` | 224 | 86.494 | 224 | 1000 | apache-2.0 | [timm](https://huggingface.co/timm/beitv2_base_patch16_224.in1k_ft_in22k_in1k) |
| `beitv2_large_patch16_224.in1k_ft_in1k` | 224 | 87.414 | 224 | 1000 | apache-2.0 | [timm](https://huggingface.co/timm/beitv2_large_patch16_224.in1k_ft_in1k) |
| `beitv2_large_patch16_224.in1k_ft_in22k` | 224 | - | - | 21841 | apache-2.0 | [timm](https://huggingface.co/timm/beitv2_large_patch16_224.in1k_ft_in22k) |
| `beitv2_large_patch16_224.in1k_ft_in22k_in1k` | 224 | 88.406 | 224 | 1000 | apache-2.0 | [timm](https://huggingface.co/timm/beitv2_large_patch16_224.in1k_ft_in22k_in1k) |

## Citation

```bibtex
@inproceedings{beit,
title={{BEiT}: {BERT} Pre-Training of Image Transformers},
author={Hangbo Bao and Li Dong and Songhao Piao and Furu Wei},
booktitle={International Conference on Learning Representations},
year={2022},
url={https://openreview.net/forum?id=p-BhZSz59o4}
}

@article{beitv2,
title={{BEiT v2}: Masked Image Modeling with Vector-Quantized Visual Tokenizers},
author={Zhiliang Peng and Li Dong and Hangbo Bao and Qixiang Ye and Furu Wei},
year={2022},
eprint={2208.06366},
archivePrefix={arXiv},
primaryClass={cs.CV}
}
```