# CaiT

**Category:** ViT · **Code:**
[`timm/models/cait.py`](https://github.com/huggingface/pytorch-image-models/blob/main/timm/models/cait.py) ·
**Architectures:** 10 · **Pretrained weights:** 10 (10 with ImageNet-1k results)

Class-Attention in Image Transformers (CaiT), deep Vision Transformers that use LayerScale and finish with
class-attention layers, which update only the class token from the patch tokens.

## Papers

* [Going deeper with Image Transformers](https://arxiv.org/abs/2103.17239)

## 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('cait_xs24_384.fb_dist_in1k', pretrained=True)
>>> data_config = timm.data.resolve_model_data_config(model)
```

## Model API support

| Feature | `Cait` |
|---|---|
| [`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 (`Cait` with `cait_xxs24_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 |
|---|---:|---:|---:|---:|
| `cait_m36_384` | 271.22 | 173.11 | 384 | 1 |
| `cait_m48_448` | 356.46 | 329.41 | 448 | 1 |
| `cait_s24_224` | 46.92 | 9.35 | 224 | 1 |
| `cait_s24_384` | 47.06 | 32.17 | 384 | 1 |
| `cait_s36_384` | 68.37 | 47.99 | 384 | 1 |
| `cait_xs24_384` | 26.67 | 19.28 | 384 | 1 |
| `cait_xxs24_224` | 11.96 | 2.53 | 224 | 1 |
| `cait_xxs24_384` | 12.03 | 9.63 | 384 | 1 |
| `cait_xxs36_224` | 17.30 | 3.77 | 224 | 1 |
| `cait_xxs36_384` | 17.37 | 14.35 | 384 | 1 |

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 |
|---|---:|---:|---:|---:|---|---|
| `cait_m36_384.fb_dist_in1k` | 384 | 86.060 | 384 | 1000 | apache-2.0 | [timm](https://huggingface.co/timm/cait_m36_384.fb_dist_in1k) |
| `cait_m48_448.fb_dist_in1k` | 448 | 86.480 | 448 | 1000 | apache-2.0 | [timm](https://huggingface.co/timm/cait_m48_448.fb_dist_in1k) |
| `cait_s24_224.fb_dist_in1k` | 224 | 83.484 | 224 | 1000 | apache-2.0 | [timm](https://huggingface.co/timm/cait_s24_224.fb_dist_in1k) |
| `cait_s24_384.fb_dist_in1k` | 384 | 85.048 | 384 | 1000 | apache-2.0 | [timm](https://huggingface.co/timm/cait_s24_384.fb_dist_in1k) |
| `cait_s36_384.fb_dist_in1k` | 384 | 85.458 | 384 | 1000 | apache-2.0 | [timm](https://huggingface.co/timm/cait_s36_384.fb_dist_in1k) |
| `cait_xs24_384.fb_dist_in1k` | 384 | 84.064 | 384 | 1000 | apache-2.0 | [timm](https://huggingface.co/timm/cait_xs24_384.fb_dist_in1k) |
| `cait_xxs24_224.fb_dist_in1k` | 224 | 78.400 | 224 | 1000 | apache-2.0 | [timm](https://huggingface.co/timm/cait_xxs24_224.fb_dist_in1k) |
| `cait_xxs24_384.fb_dist_in1k` | 384 | 80.954 | 384 | 1000 | apache-2.0 | [timm](https://huggingface.co/timm/cait_xxs24_384.fb_dist_in1k) |
| `cait_xxs36_224.fb_dist_in1k` | 224 | 79.750 | 224 | 1000 | apache-2.0 | [timm](https://huggingface.co/timm/cait_xxs36_224.fb_dist_in1k) |
| `cait_xxs36_384.fb_dist_in1k` | 384 | 82.204 | 384 | 1000 | apache-2.0 | [timm](https://huggingface.co/timm/cait_xxs36_384.fb_dist_in1k) |