{ "architectures": [ "BertForMaskedLM" ], "attention_probs_dropout_prob": 0.1, "drop_worst_after": 20000, "drop_worst_ratio": 0.2, "finetuning_task": "image_captioning", "freeze_embedding": true, "hidden_act": "gelu", "hidden_dropout_prob": 0.1, "hidden_size": 768, "img_feature_dim": 2054, "img_feature_type": "frcnn", "img_layer_norm_eps": 1e-12, "initializer_range": 0.02, "intermediate_size": 3072, "is_bert": true, "label_smoothing": 0.1, "layer_norm_eps": 1e-12, "loss_type": "sfmx", "max_position_embeddings": 512, "model_type": "bert", "num_attention_heads": 12, "num_contrast_classes": 3, "num_hidden_layers": 12, "num_labels": 2, "output_attentions": false, "output_hidden_states": false, "pad_token_id": 0, "tie_weights": true, "torchscript": false, "type_vocab_size": 2, "use_img_layernorm": 1, "vocab_size": 30522 }