lucid-dl/pvt-v2-b1

PVT v2-B1

Wang et al., 2022 — PVT v2: Improved Baselines with Pyramid Vision Transformer (arXiv:2106.13797)

Lucid port of timm/pvt_v2_b1.in1k, converted to Lucid-native safetensors.

Available weights

Tag	acc@1	acc@5	Params	GFLOPs	Size	Source
IN1K (default)	78.7	—	14.0M	—	53.46 MB	timm

Usage

import lucid.models as models
from lucid.models.weights import PVTv2B1Weights

# default tag
model = models.pvt_v2_b1_cls(pretrained=True)

# explicit tag (enum or string)
model = models.pvt_v2_b1_cls(weights=PVTv2B1Weights.IN1K)
model = models.pvt_v2_b1_cls(pretrained="IN1K")

# preprocessing travels with the weights
weights = PVTv2B1Weights.IN1K
preprocess = weights.transforms()
logits = model(preprocess(image)[None]).logits

Conversion

Converted from timm/pvt_v2_b1.in1k via python -m tools.convert_weights pvt_v2_b1 --tag IN1K. Key mapping + numerical parity verified against the source.

License

apache-2.0 — inherited from the original weights.

Citation

@article{wang2022pvtv2,
  title={PVT v2: Improved Baselines with Pyramid Vision Transformer},
  author={Wang, Wenhai and Xie, Enze and Li, Xiang and Fan, Deng-Ping and Song, Kaitao and Liang, Ding and Lu, Tong and Luo, Ping and Shao, Ling},
  journal={Computational Visual Media}, year={2022}
}