# VOLO

**Category:** ViT · **Code:**
[`timm/models/volo.py`](https://github.com/huggingface/pytorch-image-models/blob/main/timm/models/volo.py) ·
**Architectures:** 11 · **Pretrained weights:** 11 (11 with ImageNet-1k results)

Vision Outlooker (VOLO) adds outlook attention, which encodes fine-level features from local neighbourhoods, before
transformer stages that operate on coarser tokens.

## Papers

* [VOLO: Vision Outlooker for Visual Recognition](https://arxiv.org/abs/2106.13112)

## 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('volo_d1_224.sail_in1k', pretrained=True)
>>> data_config = timm.data.resolve_model_data_config(model)
```

## Model API support

| Feature | `VOLO` |
|---|---|
| [`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 (`VOLO` with `volo_d1_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 |
|---|---:|---:|---:|---:|
| `volo_d1_224` | 26.63 | 6.94 | 224 | 1 |
| `volo_d1_384` | 26.78 | 22.75 | 384 | 1 |
| `volo_d2_224` | 58.68 | 14.34 | 224 | 1 |
| `volo_d2_384` | 58.87 | 46.17 | 384 | 1 |
| `volo_d3_224` | 86.33 | 20.78 | 224 | 1 |
| `volo_d3_448` | 86.63 | 96.33 | 448 | 1 |
| `volo_d4_224` | 192.96 | 44.34 | 224 | 1 |
| `volo_d4_448` | 193.41 | 197.13 | 448 | 1 |
| `volo_d5_224` | 295.46 | 72.40 | 224 | 1 |
| `volo_d5_448` | 295.91 | 315.06 | 448 | 1 |
| `volo_d5_512` | 296.09 | 425.09 | 512 | 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 |
|---|---:|---:|---:|---:|---|---|
| `volo_d1_224.sail_in1k` | 224 | 84.148 | 224 | 1000 | apache-2.0 | [timm](https://huggingface.co/timm/volo_d1_224.sail_in1k) |
| `volo_d1_384.sail_in1k` | 384 | 85.262 | 384 | 1000 | apache-2.0 | [timm](https://huggingface.co/timm/volo_d1_384.sail_in1k) |
| `volo_d2_224.sail_in1k` | 224 | 85.204 | 224 | 1000 | apache-2.0 | [timm](https://huggingface.co/timm/volo_d2_224.sail_in1k) |
| `volo_d2_384.sail_in1k` | 384 | 86.054 | 384 | 1000 | apache-2.0 | [timm](https://huggingface.co/timm/volo_d2_384.sail_in1k) |
| `volo_d3_224.sail_in1k` | 224 | 85.440 | 224 | 1000 | apache-2.0 | [timm](https://huggingface.co/timm/volo_d3_224.sail_in1k) |
| `volo_d3_448.sail_in1k` | 448 | 86.514 | 448 | 1000 | apache-2.0 | [timm](https://huggingface.co/timm/volo_d3_448.sail_in1k) |
| `volo_d4_224.sail_in1k` | 224 | 85.868 | 224 | 1000 | apache-2.0 | [timm](https://huggingface.co/timm/volo_d4_224.sail_in1k) |
| `volo_d4_448.sail_in1k` | 448 | 86.788 | 448 | 1000 | apache-2.0 | [timm](https://huggingface.co/timm/volo_d4_448.sail_in1k) |
| `volo_d5_224.sail_in1k` | 224 | 86.076 | 224 | 1000 | apache-2.0 | [timm](https://huggingface.co/timm/volo_d5_224.sail_in1k) |
| `volo_d5_448.sail_in1k` | 448 | 86.974 | 448 | 1000 | apache-2.0 | [timm](https://huggingface.co/timm/volo_d5_448.sail_in1k) |
| `volo_d5_512.sail_in1k` | 512 | 87.068 | 512 | 1000 | apache-2.0 | [timm](https://huggingface.co/timm/volo_d5_512.sail_in1k) |