{ "_name_or_path": "roberta-base", "architectures": [ "SentenceClassificationModel" ], "attention_probs_dropout_prob": 0.1, "bos_token_id": 0, "classification_head": { "num_labels": 1, "pooling_method": "attention" }, "classifier_dropout": null, "context_layer": null, "eos_token_id": 2, "frozen_layers": [ 0, 1, 2, 3, 4, 5, 6, 7 ], "hidden_act": "gelu", "hidden_dropout_prob": 0.1, "hidden_size": 768, "initializer_range": 0.02, "intermediate_size": 3072, "layer_norm_eps": 1e-05, "max_num_sentences": 90, "max_token_length_per_sentence": 120, "max_position_embeddings": 514, "model_type": "roberta", "num_attention_heads": 12, "num_hidden_layers": 12, "pad_token_id": 1, "position_embedding_type": "absolute", "torch_dtype": "float32", "transformers_version": "4.33.2", "type_vocab_size": 1, "use_cache": true, "vocab_size": 50265 }