{ "activation": "gelu", "architectures": [ "DistilBertForSequenceClassification" ], "attention_dropout": 0.1, "dim": 768, "dropout": 0.1, "hidden_dim": 3072, "id2label": { "0": "cs", "1": "stat", "2": "astro-ph", "3": "q-bio", "4": "eess", "5": "cond-mat", "6": "math", "7": "physics", "8": "quant-ph", "9": "q-fin", "10": ":)", "11": "nlin", "12": "cmp-lg" }, "initializer_range": 0.02, "label2id": { ":)": 10, "astro-ph": 2, "cmp-lg": 12, "cond-mat": 5, "cs": 0, "eess": 4, "math": 6, "nlin": 11, "physics": 7, "q-bio": 3, "q-fin": 9, "quant-ph": 8, "stat": 1 }, "max_position_embeddings": 512, "model_type": "distilbert", "n_heads": 12, "n_layers": 6, "output_past": true, "pad_token_id": 0, "problem_type": "single_label_classification", "qa_dropout": 0.1, "seq_classif_dropout": 0.2, "sinusoidal_pos_embds": false, "tie_weights_": true, "torch_dtype": "float32", "transformers_version": "4.50.3", "vocab_size": 28996 }