---
library_name: lucid
license: bsd-3-clause
tags:
  - image-classification
  - densenet
  - lucid
datasets:
  - imagenet-1k
pipeline_tag: image-classification
model-index:
  - name: densenet-201
    results:
      - task: { type: image-classification }
        dataset: { name: ImageNet-1K, type: imagenet-1k }
        metrics:
          - { type: acc@1, value: 76.896 }
          - { type: acc@5, value: 93.37 }
---

# DenseNet-201

> Huang et al., 2017 — *Densely Connected Convolutional Networks* (arXiv:1608.06993)

[Lucid](https://github.com/ChanLumerico/lucid) port of `torchvision/DenseNet201_Weights.IMAGENET1K_V1`,
converted to Lucid-native safetensors.

## Available weights

| Tag | acc@1 | acc@5 | Params | GFLOPs | Size | Source |
|---|---|---|---|---|---|---|
| `IMAGENET1K_V1` *(default)* | 76.896 | 93.37 | 20.0M | 4.291 | 77.37 MB | torchvision |

## Usage

```python
import lucid.models as models
from lucid.models.weights import DenseNet201Weights

# default tag
model = models.densenet_201_cls(pretrained=True)

# explicit tag (enum or string)
model = models.densenet_201_cls(weights=DenseNet201Weights.IMAGENET1K_V1)
model = models.densenet_201_cls(pretrained="IMAGENET1K_V1")

# preprocessing travels with the weights
weights = DenseNet201Weights.IMAGENET1K_V1
preprocess = weights.transforms()
logits = model(preprocess(image)[None]).logits
```

## Conversion

Converted from `torchvision/DenseNet201_Weights.IMAGENET1K_V1` via
`python -m tools.convert_weights densenet_201 --tag IMAGENET1K_V1`.
Key mapping + numerical parity verified against the source.

## License

`bsd-3-clause` — inherited from the original weights.

## Citation

```
@inproceedings{huang2017densely,
  title={Densely Connected Convolutional Networks},
  author={Huang, Gao and Liu, Zhuang and Van Der Maaten, Laurens and Weinberger, Kilian Q.},
  booktitle={CVPR}, year={2017}
}
```