| { |
| "architectures": [ |
| "T5MultiClass" |
| ], |
| "classifier_dropout": 0.0, |
| "d_ff": 2048, |
| "d_kv": 64, |
| "d_model": 768, |
| "decoder_start_token_id": 0, |
| "dense_act_fn": "gelu_new", |
| "dropout_rate": 0.1, |
| "eos_token_id": 1, |
| "feed_forward_proj": "gated-gelu", |
| "gradient_checkpointing": false, |
| "id2label": { |
| "0": "acc-phys", |
| "1": "adap-org", |
| "2": "alg-geom", |
| "3": "ao-sci", |
| "4": "astro-ph", |
| "5": "atom-ph", |
| "6": "bayes-an", |
| "7": "chao-dyn", |
| "8": "chem-ph", |
| "9": "cmp-lg", |
| "10": "comp-gas", |
| "11": "cond-mat", |
| "12": "cs", |
| "13": "dg-ga", |
| "14": "econ", |
| "15": "eess", |
| "16": "funct-an", |
| "17": "gr-qc", |
| "18": "hep-ex", |
| "19": "hep-lat", |
| "20": "hep-ph", |
| "21": "hep-th", |
| "22": "math", |
| "23": "math-ph", |
| "24": "mtrl-th", |
| "25": "nlin", |
| "26": "nucl-ex", |
| "27": "nucl-th", |
| "28": "patt-sol", |
| "29": "physics", |
| "30": "plasm-ph", |
| "31": "q-alg", |
| "32": "q-bio", |
| "33": "q-fin", |
| "34": "quant-ph", |
| "35": "solv-int", |
| "36": "stat", |
| "37": "supr-con" |
| }, |
| "initializer_factor": 1.0, |
| "is_encoder_decoder": true, |
| "is_gated_act": true, |
| "label2id": { |
| "acc-phys": 0, |
| "adap-org": 1, |
| "alg-geom": 2, |
| "ao-sci": 3, |
| "astro-ph": 4, |
| "atom-ph": 5, |
| "bayes-an": 6, |
| "chao-dyn": 7, |
| "chem-ph": 8, |
| "cmp-lg": 9, |
| "comp-gas": 10, |
| "cond-mat": 11, |
| "cs": 12, |
| "dg-ga": 13, |
| "econ": 14, |
| "eess": 15, |
| "funct-an": 16, |
| "gr-qc": 17, |
| "hep-ex": 18, |
| "hep-lat": 19, |
| "hep-ph": 20, |
| "hep-th": 21, |
| "math": 22, |
| "math-ph": 23, |
| "mtrl-th": 24, |
| "nlin": 25, |
| "nucl-ex": 26, |
| "nucl-th": 27, |
| "patt-sol": 28, |
| "physics": 29, |
| "plasm-ph": 30, |
| "q-alg": 31, |
| "q-bio": 32, |
| "q-fin": 33, |
| "quant-ph": 34, |
| "solv-int": 35, |
| "stat": 36, |
| "supr-con": 37 |
| }, |
| "layer_norm_epsilon": 1e-06, |
| "model_type": "t5", |
| "num_decoder_layers": 12, |
| "num_heads": 12, |
| "num_layers": 12, |
| "output_past": true, |
| "pad_token_id": 0, |
| "relative_attention_max_distance": 128, |
| "relative_attention_num_buckets": 32, |
| "tie_word_embeddings": false, |
| "tokenizer_class": "T5Tokenizer", |
| "torch_dtype": "float32", |
| "transformers_version": "4.51.2", |
| "use_cache": true, |
| "vocab_size": 30000 |
| } |
|
|
