Datasets:

model-metadata
/

custom-vram-code

dataset_info:
  features:
    - name: model_id
      dtype: string
    - name: vram
      dtype: float64
    - name: timestamp
      dtype: string
  splits:
    - name: train
      num_bytes: 1000
      num_examples: 19
  download_size: 2161
  dataset_size: 1000
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*