seizure_eeg_train / README.md
JLB-JLB's picture
Update README.md
e83ff38
metadata
dataset_info:
  features:
    - name: image
      dtype: image
    - name: epoch
      dtype: int64
    - name: label_str
      dtype:
        class_label:
          names:
            '0': No Event
            '1': bckg
            '2': seiz
    - name: label
      dtype:
        class_label:
          names:
            '0': No Event
            '1': bckg
            '2': seiz
  splits:
    - name: train
      num_bytes: 23742147634.792
      num_examples: 814568
  download_size: 24165936927
  dataset_size: 23742147634.792
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*

Dataset Card for "seizure_eeg_train"

from datasets import load_dataset

dataset_name = "JLB-JLB/seizure_eeg_train"

dataset = load_dataset(
    dataset_name,
    split="train",
)

display(dataset)

# create train and test/val split
train_testvalid = dataset.train_test_split(test_size=0.1, shuffle=True, seed=12071998)
display(train_testvalid)

# get the number of different labels in the train, test and validation set
display(train_testvalid["train"].features["label"])
display(train_testvalid["test"].features["label"].num_classes)

# check how many labels/number of classes
num_classes = len(set(train_testvalid["train"]['label']))
labels = train_testvalid["train"].features['label']
print(um_classes, labels)


display(train_testvalid["train"][0]['image'])

More Information needed