# EfficientFormer

**Category:** Efficient/Mobile · **Code:**
[`timm/models/efficientformer.py`](https://github.com/huggingface/pytorch-image-models/blob/main/timm/models/efficientformer.py)
· **Architectures:** 3 · **Pretrained weights:** 3 (3 with ImageNet-1k results)

EfficientFormer is a vision transformer designed for mobile latency, with pooling based blocks on 4D feature maps in the
early stages and multi-head self-attention blocks in the last stage.

## Papers

* [EfficientFormer: Vision Transformers at MobileNet Speed](https://arxiv.org/abs/2206.01191)

## 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('efficientformer_l3.snap_dist_in1k', pretrained=True)
>>> data_config = timm.data.resolve_model_data_config(model)
```

## Model API support

| Feature | `EfficientFormer` |
|---|---|
| [`features_only=True` feature maps](../model_api#intermediate-features) | ✓ (strides 4, 8, 16, 32) |
| [`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 (`EfficientFormer` with `efficientformer_l1` 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 |
|---|---:|---:|---:|---:|
| `efficientformer_l1` | 12.29 | 1.30 | 224 | 1 |
| `efficientformer_l3` | 31.41 | 3.93 | 224 | 1 |
| `efficientformer_l7` | 82.23 | 10.17 | 224 | 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 |
|---|---:|---:|---:|---:|---|---|
| `efficientformer_l1.snap_dist_in1k` | 224 | 80.502 | 224 | 1000 | apache-2.0 | [timm](https://huggingface.co/timm/efficientformer_l1.snap_dist_in1k) |
| `efficientformer_l3.snap_dist_in1k` | 224 | 82.562 | 224 | 1000 | apache-2.0 | [timm](https://huggingface.co/timm/efficientformer_l3.snap_dist_in1k) |
| `efficientformer_l7.snap_dist_in1k` | 224 | 83.396 | 224 | 1000 | apache-2.0 | [timm](https://huggingface.co/timm/efficientformer_l7.snap_dist_in1k) |

## Citation

```bibtex
@article{li2022efficientformer,
  title={EfficientFormer: Vision Transformers at MobileNet Speed},
  author={Li, Yanyu and Yuan, Geng and Wen, Yang and Hu, Eric and Evangelidis, Georgios and Tulyakov,
   Sergey and Wang, Yanzhi and Ren, Jian},
  journal={arXiv preprint arXiv:2206.01191},
  year={2022}
}
```