README.md

---
library_name: lucid
license: apache-2.0
tags:
  - image-classification
  - crossvit
  - lucid
datasets:
  - imagenet-1k
pipeline_tag: image-classification
model-index:
  - name: crossvit-tiny
    results:
      - task: { type: image-classification }
        dataset: { name: ImageNet-1k, type: imagenet-1k }
        metrics:
          - { type: acc@1, value: 72.6 }
---

# CrossViT-Ti

> Chen et al., 2021 — *CrossViT: Cross-Attention Multi-Scale Vision Transformer for Image Classification* (arXiv:2103.14899)

[Lucid](https://github.com/ChanLumerico/lucid) port of `timm/crossvit_tiny_240.in1k`,
converted to Lucid-native safetensors.

## Available weights

| Tag | acc@1 | acc@5 | Params | GFLOPs | Size | Source |
|---|---|---|---|---|---|---|
| `IN1K` *(default)* | 72.6 | — | 7.0M | — | 26.79 MB | timm |

## Usage

```python
import lucid.models as models
from lucid.models.weights import CrossViTTinyWeights

# default tag
model = models.crossvit_tiny_cls(pretrained=True)

# explicit tag (enum or string)
model = models.crossvit_tiny_cls(weights=CrossViTTinyWeights.IN1K)
model = models.crossvit_tiny_cls(pretrained="IN1K")

# preprocessing travels with the weights
weights = CrossViTTinyWeights.IN1K
preprocess = weights.transforms()
logits = model(preprocess(image)[None]).logits
```

## Conversion

Converted from `timm/crossvit_tiny_240.in1k` via
`python -m tools.convert_weights crossvit_tiny --tag IN1K`.
Key mapping + numerical parity verified against the source.

## License

`apache-2.0` — inherited from the original weights.

## Citation

```
@inproceedings{chen2021crossvit,
  title={CrossViT: Cross-Attention Multi-Scale Vision Transformer for Image Classification},
  author={Chen, Chun-Fu (Richard) and Fan, Quanfu and Panda, Rameswar},
  booktitle={ICCV}, year={2021}
}
```