{ "_name_or_path": "studio-ousia/mluke-large-lite", "architectures": [ "LukeForMaskedLM" ], "attention_probs_dropout_prob": 0.1, "bert_model_name": "xlm-roberta-large", "bos_token_id": 0, "cls_entity_prediction": false, "deepspeed_transformer_layer_args": {}, "entity_emb_size": 256, "entity_vocab_size": 4, "eos_token_id": 2, "gradient_checkpointing": false, "hidden_act": "gelu", "hidden_dropout_prob": 0.1, "hidden_size": 1024, "initializer_range": 0.02, "intermediate_size": 4096, "layer_norm_eps": 1e-05, "max_position_embeddings": 514, "model_type": "luke", "num_attention_heads": 16, "num_hidden_layers": 24, "output_past": true, "pad_token_id": 1, "position_embedding_type": "absolute", "torch_dtype": "float32", "transformers_version": "4.15.0", "type_vocab_size": 1, "use_cache": true, "use_deepspeed_transformer_layer": false, "use_entity_aware_attention": false, "vocab_size": 250004 }