{ "_name_or_path": "bart-759M", "activation_dropout": 0.1, "activation_function": "gelu", "add_bias_logits": false, "add_final_layer_norm": false, "architectures": [ "BartForConditionalGeneration" ], "attention_dropout": 0.1, "bos_token_id": 101, "classif_dropout": 0.1, "classifier_dropout": 0.0, "d_model": 1024, "d_kv": 64, "decoder_attention_heads": 16, "decoder_ffn_dim": 4096, "decoder_layerdrop": 0.0, "decoder_layers": 24, "decoder_start_token_id": 102, "dropout": 0.1, "encoder_attention_heads": 16, "encoder_ffn_dim": 4096, "encoder_layerdrop": 0.0, "encoder_layers": 24, "eos_token_id": 102, "forced_eos_token_id": 102, "num_labels": 3, "init_std": 0.02, "is_encoder_decoder": true, "max_position_embeddings": 1024, "model_type": "bart", "no_repeat_ngram_size": 3, "normalize_before": false, "normalize_embedding": true, "num_beams": 4, "num_hidden_layers": 24, "pad_token_id": 0, "scale_embedding": false, "torch_dtype": "float16", "transformers_version": "4.16.0.dev0", "use_cache": true, "vocab_size": 50265 }