{ "activation_dropout": 0.0, "activation_function": "gelu", "architectures": [ "BartForConditionalGeneration" ], "attention_dropout": 0.0, "bos_token_id": 1, "classifier_dropout": 0.0, "d_model": 256, "decoder_attention_heads": 4, "decoder_ffn_dim": 1024, "decoder_layerdrop": 0, "decoder_layers": 4, "decoder_start_token_id": 2, "dropout": 0.3, "encoder_attention_heads": 4, "encoder_ffn_dim": 1024, "encoder_layerdrop": 0, "encoder_layers": 4, "eos_token_id": 2, "forced_eos_token_id": 2, "init_std": 0.02, "is_encoder_decoder": true, "max_length": 100, "max_position_embeddings": 1024, "model_type": "bart", "num_beams": 4, "num_hidden_layers": 4, "pad_token_id": 0, "scale_embedding": false, "torch_dtype": "float32", "transformers_version": "4.41.2", "unk_token_id": 3, "use_cache": true, "vocab_size": 122 }