convnext-tiny / README.md
ChanLumerico's picture
Fix usage example: import weights enum from lucid.models.weights
a7d577e verified
metadata
library_name: lucid
license: bsd-3-clause
tags:
  - image-classification
  - convnext
  - lucid
datasets:
  - imagenet-1k
pipeline_tag: image-classification
model-index:
  - name: convnext-tiny
    results:
      - task:
          type: image-classification
        dataset:
          name: ImageNet-1K
          type: imagenet-1k
        metrics:
          - type: acc@1
            value: 82.52
          - type: acc@5
            value: 96.146

ConvNeXt-Tiny

Liu et al., 2022 — A ConvNet for the 2020s (arXiv:2201.03545)

Lucid port of torchvision/ConvNeXt_Tiny_Weights.IMAGENET1K_V1, converted to Lucid-native safetensors.

Available weights

Tag acc@1 acc@5 Params GFLOPs Size Source
IMAGENET1K_V1 (default) 82.52 96.146 28.6M 4.456 109.07 MB torchvision

Usage

import lucid.models as models
from lucid.models.weights import ConvNeXtTinyWeights

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

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

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

Conversion

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

License

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

Citation

@inproceedings{liu2022convnet,
  title={A ConvNet for the 2020s},
  author={Liu, Zhuang and Mao, Hanzi and Wu, Chao-Yuan and Feichtenhofer, Christoph and Darrell, Trevor and Xie, Saining},
  booktitle={CVPR}, year={2022}
}