imagenet_pi

  • Mô tả :

ImageNet-PI là phiên bản được dán nhãn lại của bộ dữ liệu ImageNet ILSVRC2012 tiêu chuẩn, trong đó các nhãn được cung cấp bởi một tập hợp gồm 16 mạng thần kinh sâu với các kiến ​​trúc khác nhau được đào tạo trước trên ILSVRC2012 tiêu chuẩn. Cụ thể, các mô hình được đào tạo trước được tải xuống từ tf.keras.applications.

Ngoài các nhãn mới, ImageNet-PI cũng cung cấp siêu dữ liệu về quy trình chú thích dưới dạng lời khai của các mô hình trên nhãn của chúng và thông tin bổ sung về từng mô hình.

Để biết thêm thông tin, hãy xem: ImageNet-PI

  • Trang chủ : https://github.com/google-research-datasets/imagenet_pi/

  • Mã nguồn : tfds.datasets.imagenet_pi.Builder

  • Phiên bản :

    • 1.0.0 (mặc định): Bản phát hành đầu tiên.
  • Kích thước tải xuống : Unknown size

  • Kích thước tập dữ liệu : Unknown size

  • Hướng dẫn tải xuống thủ công : Bộ dữ liệu này yêu cầu bạn tải xuống dữ liệu nguồn theo cách thủ công vào download_config.manual_dir (mặc định là ~/tensorflow_datasets/downloads/manual/ ):
    manual_dir phải chứa hai tệp: ILSVRC2012_img_train.tar và ILSVRC2012_img_val.tar. Bạn cần đăng ký trên http://www.image-net.org/download-images để nhận liên kết tải xuống tập dữ liệu.

  • Tự động lưu vào bộ nhớ đệm ( tài liệu ): Không xác định

  • Chia tách :

Tách ra ví dụ
  • Cấu trúc tính năng :
FeaturesDict({
    'annotator_confidences': Tensor(shape=(16,), dtype=float32),
    'annotator_labels': Tensor(shape=(16,), dtype=int64),
    'clean_label': ClassLabel(shape=(), dtype=int64, num_classes=1000),
    'file_name': Text(shape=(), dtype=string),
    'image': Image(shape=(None, None, 3), dtype=uint8),
})
  • Tài liệu tính năng :
Tính năng Lớp học Hình dạng Dtype Sự miêu tả
Tính năngDict
annotator_confids tenxơ (16,) phao32
chú thích_nhãn tenxơ (16,) int64
clean_label LớpNhãn int64
tên_tệp Chữ sợi dây
hình ảnh Hình ảnh (Không, Không có, 3) uint8
@inproceedings{tram,
  author    = {Mark Collier and
               Rodolphe Jenatton and
               Effrosyni Kokiopoulou and
               Jesse Berent},
  editor    = {Kamalika Chaudhuri and
               Stefanie Jegelka and
               Le Song and
               Csaba Szepesv{\'{a} }ri and
               Gang Niu and
               Sivan Sabato},
  title     = {Transfer and Marginalize: Explaining Away Label Noise with Privileged
               Information},
  booktitle = {International Conference on Machine Learning, {ICML} 2022, 17-23 July
               2022, Baltimore, Maryland, {USA} },
  series    = {Proceedings of Machine Learning Research},
  volume    = {162},
  pages     = {4219--4237},
  publisher = { {PMLR} },
  year      = {2022},
  url       = {https://proceedings.mlr.press/v162/collier22a.html},
}
@article{ILSVRC15,
Author = {Olga Russakovsky and Jia Deng and Hao Su and Jonathan Krause and Sanjeev Satheesh and Sean Ma and Zhiheng Huang and Andrej Karpathy and Aditya Khosla and Michael Bernstein and Alexander C. Berg and Li Fei-Fei},
Title = { {ImageNet Large Scale Visual Recognition Challenge} },
Year = {2015},
journal   = {International Journal of Computer Vision (IJCV)},
doi = {10.1007/s11263-015-0816-y},
volume={115},
number={3},
pages={211-252}
}
,

  • Mô tả :

ImageNet-PI là phiên bản được dán nhãn lại của bộ dữ liệu ImageNet ILSVRC2012 tiêu chuẩn, trong đó các nhãn được cung cấp bởi một tập hợp gồm 16 mạng thần kinh sâu với các kiến ​​trúc khác nhau được đào tạo trước trên ILSVRC2012 tiêu chuẩn. Cụ thể, các mô hình được đào tạo trước được tải xuống từ tf.keras.applications.

Ngoài các nhãn mới, ImageNet-PI cũng cung cấp siêu dữ liệu về quy trình chú thích dưới dạng lời khai của các mô hình trên nhãn của chúng và thông tin bổ sung về từng mô hình.

Để biết thêm thông tin, hãy xem: ImageNet-PI

  • Trang chủ : https://github.com/google-research-datasets/imagenet_pi/

  • Mã nguồn : tfds.datasets.imagenet_pi.Builder

  • Phiên bản :

    • 1.0.0 (mặc định): Bản phát hành đầu tiên.
  • Kích thước tải xuống : Unknown size

  • Kích thước tập dữ liệu : Unknown size

  • Hướng dẫn tải xuống thủ công : Bộ dữ liệu này yêu cầu bạn tải xuống dữ liệu nguồn theo cách thủ công vào download_config.manual_dir (mặc định là ~/tensorflow_datasets/downloads/manual/ ):
    manual_dir phải chứa hai tệp: ILSVRC2012_img_train.tar và ILSVRC2012_img_val.tar. Bạn cần đăng ký trên http://www.image-net.org/download-images để nhận liên kết tải xuống tập dữ liệu.

  • Tự động lưu vào bộ nhớ cache ( tài liệu ): Không xác định

  • Chia tách :

Tách ra ví dụ
  • Cấu trúc tính năng :
FeaturesDict({
    'annotator_confidences': Tensor(shape=(16,), dtype=float32),
    'annotator_labels': Tensor(shape=(16,), dtype=int64),
    'clean_label': ClassLabel(shape=(), dtype=int64, num_classes=1000),
    'file_name': Text(shape=(), dtype=string),
    'image': Image(shape=(None, None, 3), dtype=uint8),
})
  • Tài liệu tính năng :
Tính năng Lớp học Hình dạng Dtype Sự miêu tả
Tính năngDict
annotator_confids tenxơ (16,) phao32
chú thích_nhãn tenxơ (16,) int64
clean_label LớpNhãn int64
tên_tệp Chữ sợi dây
hình ảnh Hình ảnh (Không, Không có, 3) uint8
@inproceedings{tram,
  author    = {Mark Collier and
               Rodolphe Jenatton and
               Effrosyni Kokiopoulou and
               Jesse Berent},
  editor    = {Kamalika Chaudhuri and
               Stefanie Jegelka and
               Le Song and
               Csaba Szepesv{\'{a} }ri and
               Gang Niu and
               Sivan Sabato},
  title     = {Transfer and Marginalize: Explaining Away Label Noise with Privileged
               Information},
  booktitle = {International Conference on Machine Learning, {ICML} 2022, 17-23 July
               2022, Baltimore, Maryland, {USA} },
  series    = {Proceedings of Machine Learning Research},
  volume    = {162},
  pages     = {4219--4237},
  publisher = { {PMLR} },
  year      = {2022},
  url       = {https://proceedings.mlr.press/v162/collier22a.html},
}
@article{ILSVRC15,
Author = {Olga Russakovsky and Jia Deng and Hao Su and Jonathan Krause and Sanjeev Satheesh and Sean Ma and Zhiheng Huang and Andrej Karpathy and Aditya Khosla and Michael Bernstein and Alexander C. Berg and Li Fei-Fei},
Title = { {ImageNet Large Scale Visual Recognition Challenge} },
Year = {2015},
journal   = {International Journal of Computer Vision (IJCV)},
doi = {10.1007/s11263-015-0816-y},
volume={115},
number={3},
pages={211-252}
}