Cerebral Palsy (CP) is a neurological disorder that is usually caused by mechanical damage to the brain at or around birth. One of the main symptoms of CP is impaired motor control and development of walking skills. In clinical practice in the US child’s gait is often recorded with a camera and the recording can be used by a doctor for assessment. In this dataset we provide body landmarks (such as positions of knees, elbow, toes etc.) extracted using OpenPose from such videos and corresponding labels derived from optical motion capture or annotated by physicians, including speed, cadence, GDI, knee flexion at max extension, surgerical decisions. This dataset can be used, for example, for designing algorithms that estimate severity of a disease from a single camera video.