模型:
facebook/convnextv2-large-22k-224
ConvNeXt V2 模型使用 FCMAE 框架在 ImageNet-22K 数据集上进行了预训练,并在分辨率为 224x224 的情况下进行了微调。该模型由 Woo 等人在论文 ConvNeXt V2: Co-designing and Scaling ConvNets with Masked Autoencoders 中提出,并于 this repository 首次发布。
免责声明:发布 ConvNeXT V2 模型的团队未为该模型编写模型卡片,因此本模型卡片由 Hugging Face 团队编写。
ConvNeXt V2 是一个纯卷积模型(ConvNet),它引入了完全卷积的掩码自编码器框架(FCMAE)和新的全局响应归一化(GRN)层到 ConvNeXt。ConvNeXt V2 显著提高了纯 ConvNet 在各种识别基准上的性能。
您可以使用原始模型进行图像分类。查看 model hub 了解您感兴趣的任务上的预训练版本。
这是如何使用此模型将 COCO 2017 数据集中的图像分类为 1000 个 ImageNet 类别之一的方法:
from transformers import AutoImageProcessor, ConvNextV2ForImageClassification
import torch
from datasets import load_dataset
dataset = load_dataset("huggingface/cats-image")
image = dataset["test"]["image"][0]
preprocessor = AutoImageProcessor.from_pretrained("facebook/convnextv2-large-22k-224")
model = ConvNextV2ForImageClassification.from_pretrained("facebook/convnextv2-large-22k-224")
inputs = preprocessor(image, return_tensors="pt")
with torch.no_grad():
logits = model(**inputs).logits
# model predicts one of the 1000 ImageNet classes
predicted_label = logits.argmax(-1).item()
print(model.config.id2label[predicted_label]),
关于更多代码示例,请参考 documentation 。
@article{DBLP:journals/corr/abs-2301-00808,
author = {Sanghyun Woo and
Shoubhik Debnath and
Ronghang Hu and
Xinlei Chen and
Zhuang Liu and
In So Kweon and
Saining Xie},
title = {ConvNeXt {V2:} Co-designing and Scaling ConvNets with Masked Autoencoders},
journal = {CoRR},
volume = {abs/2301.00808},
year = {2023},
url = {https://doi.org/10.48550/arXiv.2301.00808},
doi = {10.48550/arXiv.2301.00808},
eprinttype = {arXiv},
eprint = {2301.00808},
timestamp = {Tue, 10 Jan 2023 15:10:12 +0100},
biburl = {https://dblp.org/rec/journals/corr/abs-2301-00808.bib},
bibsource = {dblp computer science bibliography, https://dblp.org}
}