HOW TO DEPLOY

Model to Fast API and Docker

docker build -t seetrue_ai ./fast_server
docker run --rm -p 8080:8080 -d seetrue_ai

Payload

Request Payload to /predict

{
  "timestamp": 5,
  "gazepoint_x": 0.7757,
  "gazepoint_y": 0.1935,
  "pupil_area_right_sq_mm": 0.12,
  "pupil_area_left_sq_mm": 0.2,
  "eye_event": "FE",
  "euclidean_distance": 0.0132,
  "prev_euclidean_distance": null
}

Expected Response

{
    "predictions": [
        0.826665997505188,
        0.1633332520723343,
        0.009999999776482582
    ],
    "label_classes": [
        "1",
        "3",
        "2"
    ],
    "prev_euclidean_distance": 0.0132
}