MIMIC‑CXR Dataset

Dataset Information

Features

• image: image data (image type).
• findings: textual description of findings (string type).
• impression: textual overall impression (string type).

Data Splits

• train: training set, 30,633 samples, 800,678,886 bytes.

Data Size

• Download size: 792,886,513 bytes
• Dataset size: 800,678,886 bytes

Configuration

• config_name: default
  • data_files:
    • split: train
      • path: data/train-*