config.json

{
    "text_encoder": {
        "tokenizer_class": "bert",
        "model_type": "bert",
        "dim": 768,
        "context_dim": 384,
        "vocab_size": 30522,
        "padding_idx": 0,
        "num_layers": 4,
        "num_heads": 12,
        "embedding_dim": 256,
        "multimodal_layers_ids": [
            2,
            3
        ],
        "head_one_neuron": false,
        "pooling": "cls",
        "max_position_embeddings": 64,
        "dropout_prob": 0.1
    },
    "image_encoder": {
        "dim": 384,
        "patch_size": 16,
        "image_size": 224,
        "num_layers": 12,
        "num_heads": 6,
        "embedding_dim": 256,
        "normalization_means": [0.48145466, 0.4578275, 0.40821073],
        "normalization_deviations": [0.26862954, 0.26130258, 0.27577711],
        "pooling": "cls"
    }
}