Datasets:
timm
/

Modalities:
Image
Formats:
parquet
ArXiv:
Libraries:
Datasets
Dask
License:
mini-imagenet / README.md
rwightman's picture
rwightman HF staff
Update README.md
bd8779f verified
metadata
dataset_info:
  features:
    - name: image
      dtype: image
    - name: label
      dtype:
        class_label:
          names:
            '0': n01532829
            '1': n01558993
            '2': n01704323
            '3': n01749939
            '4': n01770081
            '5': n01843383
            '6': n01855672
            '7': n01910747
            '8': n01930112
            '9': n01981276
            '10': n02074367
            '11': n02089867
            '12': n02091244
            '13': n02091831
            '14': n02099601
            '15': n02101006
            '16': n02105505
            '17': n02108089
            '18': n02108551
            '19': n02108915
            '20': n02110063
            '21': n02110341
            '22': n02111277
            '23': n02113712
            '24': n02114548
            '25': n02116738
            '26': n02120079
            '27': n02129165
            '28': n02138441
            '29': n02165456
            '30': n02174001
            '31': n02219486
            '32': n02443484
            '33': n02457408
            '34': n02606052
            '35': n02687172
            '36': n02747177
            '37': n02795169
            '38': n02823428
            '39': n02871525
            '40': n02950826
            '41': n02966193
            '42': n02971356
            '43': n02981792
            '44': n03017168
            '45': n03047690
            '46': n03062245
            '47': n03075370
            '48': n03127925
            '49': n03146219
            '50': n03207743
            '51': n03220513
            '52': n03272010
            '53': n03337140
            '54': n03347037
            '55': n03400231
            '56': n03417042
            '57': n03476684
            '58': n03527444
            '59': n03535780
            '60': n03544143
            '61': n03584254
            '62': n03676483
            '63': n03770439
            '64': n03773504
            '65': n03775546
            '66': n03838899
            '67': n03854065
            '68': n03888605
            '69': n03908618
            '70': n03924679
            '71': n03980874
            '72': n03998194
            '73': n04067472
            '74': n04146614
            '75': n04149813
            '76': n04243546
            '77': n04251144
            '78': n04258138
            '79': n04275548
            '80': n04296562
            '81': n04389033
            '82': n04418357
            '83': n04435653
            '84': n04443257
            '85': n04509417
            '86': n04515003
            '87': n04522168
            '88': n04596742
            '89': n04604644
            '90': n04612504
            '91': n06794110
            '92': n07584110
            '93': n07613480
            '94': n07697537
            '95': n07747607
            '96': n09246464
            '97': n09256479
            '98': n13054560
            '99': n13133613
  splits:
    - name: train
      num_bytes: 6284840508
      num_examples: 50000
    - name: validation
      num_bytes: 1286953696
      num_examples: 10000
    - name: test
      num_bytes: 670707560
      num_examples: 5000
  download_size: 7433461683
  dataset_size: 8242501764
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: validation
        path: data/validation-*
      - split: test
        path: data/test-*
license: other
license_name: imagenet
license_link: https://www.image-net.org/download.php
task_categories:
  - image-classification
pretty_name: Mini-ImageNet
size_categories:
  - 10K<n<100K

Dataset Description

A mini version of ImageNet-1k with 100 of 1000 classes present.

Unlike some 'mini' variants this one includes the original images at their original sizes. Many such subsets downsample to 84x84 or other smaller resolutions.

Data Splits

Train

  • 50000 samples from ImageNet-1k train split

Validation

  • 10000 samples from ImageNet-1k train split

Test

  • 5000 samples from ImageNet-1k validation split (all 50 samples per class)

Usage

This dataset is good for testing hparams and models in timm

Train

python train.py --dataset hfds/timm/mini-imagenet --model resnet50 --amp --num-classes 100

Citation Information

For the specific instance of this mini variant I am not sure what the origin is. It is different from commonly referenced Vinyales et al.,2016 as it doesn't match the classes / splits.

Train & validation splits match train & test of https://www.kaggle.com/datasets/ctrnngtrung/miniimagenet ... it is not clear where that originated though.

Original ImageNet citation:

@article{imagenet15russakovsky,
    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}
}