| { |
| "_name_or_path": "/mnt/data4/luyiheng/BARTfinetune/lr_5e-05/checkpoint-195", |
| "_num_labels": 3, |
| "activation_dropout": 0.0, |
| "activation_function": "gelu", |
| "add_final_layer_norm": false, |
| "architectures": [ |
| "BartForSequenceClassification" |
| ], |
| "attention_dropout": 0.0, |
| "bos_token_id": 0, |
| "classif_dropout": 0.0, |
| "classifier_dropout": 0.0, |
| "d_model": 1024, |
| "decoder_attention_heads": 16, |
| "decoder_ffn_dim": 4096, |
| "decoder_layerdrop": 0.0, |
| "decoder_layers": 12, |
| "decoder_start_token_id": 2, |
| "dropout": 0.1, |
| "encoder_attention_heads": 16, |
| "encoder_ffn_dim": 4096, |
| "encoder_layerdrop": 0.0, |
| "encoder_layers": 12, |
| "eos_token_id": 2, |
| "forced_eos_token_id": 2, |
| "gradient_checkpointing": false, |
| "id2label": { |
| "0": "Accessibility", |
| "1": "Architectures", |
| "2": "Artificial intelligence", |
| "3": "Arts and humanities", |
| "4": "Collaborative and social computing", |
| "5": "Communication hardware, interfaces and storage", |
| "6": "Computational complexity and cryptography", |
| "7": "Computers in other domains", |
| "8": "Continuous mathematics", |
| "9": "Cross-computing tools and techniques", |
| "10": "Cryptography", |
| "11": "Data management systems", |
| "12": "Database and storage security", |
| "13": "Dependable and fault-tolerant systems and networks", |
| "14": "Design and analysis of algorithms", |
| "15": "Discrete mathematics", |
| "16": "Distributed computing methodologies", |
| "17": "Document management and text processing", |
| "18": "Document types", |
| "19": "Education", |
| "20": "Electronic commerce", |
| "21": "Electronic design automation", |
| "22": "Embedded and cyber-physical systems", |
| "23": "Emerging technologies", |
| "24": "Enterprise computing", |
| "25": "Formal languages and automata theory", |
| "26": "Formal methods and theory of security", |
| "27": "Hardware test", |
| "28": "Hardware validation", |
| "29": "Human and societal aspects of security and privacy", |
| "30": "Human computer interaction (HCI)", |
| "31": "Information retrieval", |
| "32": "Information storage systems", |
| "33": "Information systems applications", |
| "34": "Information theory", |
| "35": "Integrated circuits", |
| "36": "Interaction design", |
| "37": "Intrusion/anomaly detection and malware mitigation", |
| "38": "Logic", |
| "39": "Machine learning", |
| "40": "Mathematical analysis", |
| "41": "Mathematical software", |
| "42": "Modeling and simulation", |
| "43": "Models of computation", |
| "44": "Network algorithms", |
| "45": "Network architectures", |
| "46": "Network components", |
| "47": "Network performance evaluation", |
| "48": "Network properties", |
| "49": "Network protocols", |
| "50": "Network security", |
| "51": "Network services", |
| "52": "Network types", |
| "53": "Parallel computing methodologies", |
| "54": "Power and energy", |
| "55": "Probability and statistics", |
| "56": "Randomness, geometry and discrete structures", |
| "57": "Real-time systems", |
| "58": "Robustness", |
| "59": "Security in hardware", |
| "60": "Security services", |
| "61": "Semantics and reasoning", |
| "62": "Software and application security", |
| "63": "Symbolic and algebraic manipulation", |
| "64": "Systems security", |
| "65": "Theory and algorithms for application domains", |
| "66": "Ubiquitous and mobile computing", |
| "67": "Very large scale integration design", |
| "68": "Visualization", |
| "69": "World Wide Web" |
| }, |
| "init_std": 0.02, |
| "is_encoder_decoder": true, |
| "label2id": { |
| "Accessibility": 0, |
| "Architectures": 1, |
| "Artificial intelligence": 2, |
| "Arts and humanities": 3, |
| "Collaborative and social computing": 4, |
| "Communication hardware, interfaces and storage": 5, |
| "Computational complexity and cryptography": 6, |
| "Computers in other domains": 7, |
| "Continuous mathematics": 8, |
| "Cross-computing tools and techniques": 9, |
| "Cryptography": 10, |
| "Data management systems": 11, |
| "Database and storage security": 12, |
| "Dependable and fault-tolerant systems and networks": 13, |
| "Design and analysis of algorithms": 14, |
| "Discrete mathematics": 15, |
| "Distributed computing methodologies": 16, |
| "Document management and text processing": 17, |
| "Document types": 18, |
| "Education": 19, |
| "Electronic commerce": 20, |
| "Electronic design automation": 21, |
| "Embedded and cyber-physical systems": 22, |
| "Emerging technologies": 23, |
| "Enterprise computing": 24, |
| "Formal languages and automata theory": 25, |
| "Formal methods and theory of security": 26, |
| "Hardware test": 27, |
| "Hardware validation": 28, |
| "Human and societal aspects of security and privacy": 29, |
| "Human computer interaction (HCI)": 30, |
| "Information retrieval": 31, |
| "Information storage systems": 32, |
| "Information systems applications": 33, |
| "Information theory": 34, |
| "Integrated circuits": 35, |
| "Interaction design": 36, |
| "Intrusion/anomaly detection and malware mitigation": 37, |
| "Logic": 38, |
| "Machine learning": 39, |
| "Mathematical analysis": 40, |
| "Mathematical software": 41, |
| "Modeling and simulation": 42, |
| "Models of computation": 43, |
| "Network algorithms": 44, |
| "Network architectures": 45, |
| "Network components": 46, |
| "Network performance evaluation": 47, |
| "Network properties": 48, |
| "Network protocols": 49, |
| "Network security": 50, |
| "Network services": 51, |
| "Network types": 52, |
| "Parallel computing methodologies": 53, |
| "Power and energy": 54, |
| "Probability and statistics": 55, |
| "Randomness, geometry and discrete structures": 56, |
| "Real-time systems": 57, |
| "Robustness": 58, |
| "Security in hardware": 59, |
| "Security services": 60, |
| "Semantics and reasoning": 61, |
| "Software and application security": 62, |
| "Symbolic and algebraic manipulation": 63, |
| "Systems security": 64, |
| "Theory and algorithms for application domains": 65, |
| "Ubiquitous and mobile computing": 66, |
| "Very large scale integration design": 67, |
| "Visualization": 68, |
| "World Wide Web": 69 |
| }, |
| "max_position_embeddings": 1024, |
| "model_type": "bart", |
| "normalize_before": false, |
| "num_hidden_layers": 12, |
| "output_past": false, |
| "pad_token_id": 1, |
| "problem_type": "single_label_classification", |
| "scale_embedding": false, |
| "torch_dtype": "float32", |
| "transformers_version": "4.49.0", |
| "use_cache": true, |
| "vocab_size": 50265 |
| } |
|
|
