{
    "model_type": "cvt",
    "architectures": [
        "CvTForImageClassification"
    ],
    "paper": {
        "title": "CornViT: A Multi-Stage Convolutional Vision Transformer Framework for Hierarchical Corn Kernel Analysis",
        "year": 2025,
        "doi": "https://doi.org/10.3390/computers15010002"
    },
    "pipeline_tag": "image-classification",
    "library_name": "keras",
    "framework": "pytorch",
    "image_size": 384,
    "num_channels": 3,
    "pretrained": "imagenet-22k",
    "backbone": "cvt-13",
    "hierarchical_pipeline": true,
    "stages": [
        {
            "stage": 1,
            "name": "purity_detection",
            "labels": {
                "0": "pure",
                "1": "impure"
            },
            "num_labels": 2,
            "accuracy": 0.938
        },
        {
            "stage": 2,
            "name": "shape_classification",
            "labels": {
                "0": "flat",
                "1": "round"
            },
            "num_labels": 2,
            "accuracy": 0.941
        },
        {
            "stage": 3,
            "name": "embryo_orientation",
            "labels": {
                "0": "up",
                "1": "down"
            },
            "num_labels": 2,
            "accuracy": 0.911
        }
    ],
    "training_framework": "pytorch"
}