| { | |
| "_name_or_path": "./", | |
| "activation_dropout": 0.0, | |
| "activation_function": "relu", | |
| "add_bias_logits": false, | |
| "add_final_layer_norm": false, | |
| "share_encoder_decoder_embeddings": true, | |
| "encoder_normalize_embedding": false, | |
| "decoder_normalize_embedding": false, | |
| "architectures": [ | |
| "BartModel" | |
| ], | |
| "auto_map": { | |
| "AutoConfig": "configuration_bat.BartConfig", | |
| "AutoModel": "modeling_bat.BartForConditionalGeneration", | |
| "AutoModelForSeq2SeqLM": "modeling_bat.BartForConditionalGeneration" | |
| }, | |
| "attention_dropout": 0.0, | |
| "bos_token_id": 0, | |
| "d_model": 1024, | |
| "decoder_attention_heads": 16, | |
| "decoder_ffn_dim": 4096, | |
| "decoder_layerdrop": 0.05, | |
| "decoder_layers": 6, | |
| "decoder_start_token_id": 2, | |
| "dropout": 0.1, | |
| "early_stopping": true, | |
| "encoder_attention_heads": 16, | |
| "encoder_ffn_dim": 4096, | |
| "encoder_layerdrop": 0.05, | |
| "encoder_layers": 6, | |
| "eos_token_id": 2, | |
| "gradient_checkpointing": false, | |
| "init_std": 0.02, | |
| "is_encoder_decoder": true, | |
| "max_length": 1024, | |
| "max_position_embeddings": 1024, | |
| "model_type": "bart", | |
| "num_beams": 5, | |
| "num_hidden_layers": 12, | |
| "pad_token_id": 1, | |
| "scale_embedding": true, | |
| "transformers_version": "4.4.0.dev0", | |
| "use_cache": true, | |
| "forced_bos_token_id": 64870, | |
| "vocab_size": 64871 | |
| } |