config.json

{
    "model_type": "tvp",
    "alpha": 1.0,
    "beta": 0.1,
    "vp_type": "framepad",
    "vp_apply": "replace",
    "max_img_size": 448,
    "pad_size": 96,
    "num_frm": 48,
    "attention_probs_dropout_prob": 0.1,
    "hidden_size": 768,
    "classifier": "mlp",
    "classifier_dropout": null,
    "cls_hidden_scale": 2,
    "hidden_act": "gelu",
    "hidden_dropout_prob": 0.1,
    "initializer_range": 0.02,
    "intermediate_size": 3072,
    "layer_norm_eps": 1e-12,
    "max_grid_col_position_embeddings": 100,
    "max_grid_row_position_embeddings": 100,
    "max_position_embeddings": 512,
    "num_attention_heads": 12,
    "num_hidden_layers": 12,
    "pad_token_id": 0,
    "position_embedding_type": "absolute",
    "transformers_version": "4.29.2",
    "type_vocab_size": 2,
    "use_cache": true,
    "vocab_size": 30522,
    "max_text_length": 100,
    "vision_config": {
        "model_type": "tvp_vision_model",
        "input_format": "BGR",
        "features": ["res5"],
        "resnets_depth": 50,
        "resnets_num_groups": 1,
        "resnets_width_per_group": 64,
        "resnets_stem_input_channels": 3,
        "resnets_stem_out_channels": 64,
        "resnets_res_out_channels": 256,
        "resnets_res_dilation": 1,
        "backbone_freeze_at": 2,
        "grid_encoder_conv_input_size": 2048,
        "grid_encoder_conv_output_size": 768
    }
}