# iFormer

**Category:** Efficient/Mobile · **Code:**
[`timm/models/iformer.py`](https://github.com/huggingface/pytorch-image-models/blob/main/timm/models/iformer.py) ·
**Architectures:** 9 · **Pretrained weights:** 8 (0 with ImageNet-1k results)

iFormer integrates a fast convolutional network with single-head modulation attention (SHMA) in the later stages, for
mobile applications.

## Papers

* [iFormer: Integrating ConvNet and Transformer for Mobile Application](https://arxiv.org/abs/2501.15369)

## 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('iformer_h.in1k', pretrained=True)
>>> data_config = timm.data.resolve_model_data_config(model)
```

## Model API support

| Feature | `IFormer` |
|---|---|
| [`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 (`IFormer` with `iformer_t` 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 |
|---|---:|---:|---:|---:|
| `iformer_h` | 99.60 | - | 224 | 1 |
| `iformer_l` | 14.78 | - | 224 | 1 |
| `iformer_l2` | 24.52 | - | 224 | 0 |
| `iformer_l2_distilled` | 25.03 | - | 224 | 1 |
| `iformer_l_distilled` | 15.17 | - | 224 | 1 |
| `iformer_m` | 8.91 | - | 224 | 1 |
| `iformer_m_distilled` | 9.29 | - | 224 | 1 |
| `iformer_s` | 6.56 | - | 224 | 1 |
| `iformer_t` | 2.89 | - | 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 |
|---|---:|---:|---:|---:|---|---|
| `iformer_h.in1k` | 224 | - | - | 1000 | mit | [timm](https://huggingface.co/timm/iformer_h.in1k) |
| `iformer_l.in1k` | 224 | - | - | 1000 | mit | [timm](https://huggingface.co/timm/iformer_l.in1k) |
| `iformer_l2_distilled.in1k` | 224 | - | - | 1000 | mit | [timm](https://huggingface.co/timm/iformer_l2_distilled.in1k) |
| `iformer_l_distilled.in1k` | 224 | - | - | 1000 | mit | [timm](https://huggingface.co/timm/iformer_l_distilled.in1k) |
| `iformer_m.in1k` | 224 | - | - | 1000 | mit | [timm](https://huggingface.co/timm/iformer_m.in1k) |
| `iformer_m_distilled.in1k` | 224 | - | - | 1000 | mit | [timm](https://huggingface.co/timm/iformer_m_distilled.in1k) |
| `iformer_s.in1k` | 224 | - | - | 1000 | mit | [timm](https://huggingface.co/timm/iformer_s.in1k) |
| `iformer_t.in1k` | 224 | - | - | 1000 | mit | [timm](https://huggingface.co/timm/iformer_t.in1k) |

## Citation

```bibtex
@inproceedings{zheng2025iformer,
  title={iformer: Integrating convnet and transformer for mobile application},
  author={Zheng, Chuanyang},
  booktitle={International Conference on Learning Representations},
  volume={2025},
  pages={22947--22961},
  year={2025}
}
```