{ "architectures": [ "ModernBertForSequenceClassification" ], "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 50281, "classifier_activation": "gelu", "classifier_bias": false, "classifier_dropout": 0.0, "classifier_pooling": "mean", "cls_token_id": 50281, "decoder_bias": true, "deterministic_flash_attn": false, "dtype": "float32", "embedding_dropout": 0.0, "eos_token_id": 50282, "global_attn_every_n_layers": 3, "gradient_checkpointing": false, "hidden_activation": "gelu", "hidden_size": 768, "id2label": { "0": "abstract_noun", "1": "additive_formal", "2": "agent_demoted", "3": "agentless_passive", "4": "alliteration", "5": "anaphora", "6": "antithesis", "7": "aside", "8": "assonance", "9": "asyndeton", "10": "audience_response", "11": "categorical_statement", "12": "causal_chain", "13": "causal_explicit", "14": "citation", "15": "conceptual_metaphor", "16": "concessive", "17": "concessive_connector", "18": "conditional", "19": "conflict_frame", "20": "contrastive", "21": "cross_reference", "22": "definitional_move", "23": "discourse_formula", "24": "dramatic_pause", "25": "embodied_action", "26": "enumeration", "27": "epistemic_hedge", "28": "epistrophe", "29": "epithet", "30": "everyday_example", "31": "evidential", "32": "footnote_reference", "33": "imperative", "34": "inclusive_we", "35": "institutional_subject", "36": "intensifier_doubling", "37": "lexical_repetition", "38": "list_structure", "39": "metadiscourse", "40": "methodological_framing", "41": "named_individual", "42": "nested_clauses", "43": "nominalization", "44": "objectifying_stance", "45": "parallelism", "46": "phatic_check", "47": "phatic_filler", "48": "polysyndeton", "49": "probability", "50": "proverb", "51": "qualified_assertion", "52": "refrain", "53": "relative_chain", "54": "religious_formula", "55": "rhetorical_question", "56": "rhyme", "57": "rhythm", "58": "second_person", "59": "self_correction", "60": "sensory_detail", "61": "simple_conjunction", "62": "specific_place", "63": "technical_abbreviation", "64": "technical_term", "65": "temporal_anchor", "66": "temporal_embedding", "67": "third_person_reference", "68": "tricolon", "69": "us_them", "70": "vocative" }, "initializer_cutoff_factor": 2.0, "initializer_range": 0.02, "intermediate_size": 1152, "label2id": { "abstract_noun": 0, "additive_formal": 1, "agent_demoted": 2, "agentless_passive": 3, "alliteration": 4, "anaphora": 5, "antithesis": 6, "aside": 7, "assonance": 8, "asyndeton": 9, "audience_response": 10, "categorical_statement": 11, "causal_chain": 12, "causal_explicit": 13, "citation": 14, "conceptual_metaphor": 15, "concessive": 16, "concessive_connector": 17, "conditional": 18, "conflict_frame": 19, "contrastive": 20, "cross_reference": 21, "definitional_move": 22, "discourse_formula": 23, "dramatic_pause": 24, "embodied_action": 25, "enumeration": 26, "epistemic_hedge": 27, "epistrophe": 28, "epithet": 29, "everyday_example": 30, "evidential": 31, "footnote_reference": 32, "imperative": 33, "inclusive_we": 34, "institutional_subject": 35, "intensifier_doubling": 36, "lexical_repetition": 37, "list_structure": 38, "metadiscourse": 39, "methodological_framing": 40, "named_individual": 41, "nested_clauses": 42, "nominalization": 43, "objectifying_stance": 44, "parallelism": 45, "phatic_check": 46, "phatic_filler": 47, "polysyndeton": 48, "probability": 49, "proverb": 50, "qualified_assertion": 51, "refrain": 52, "relative_chain": 53, "religious_formula": 54, "rhetorical_question": 55, "rhyme": 56, "rhythm": 57, "second_person": 58, "self_correction": 59, "sensory_detail": 60, "simple_conjunction": 61, "specific_place": 62, "technical_abbreviation": 63, "technical_term": 64, "temporal_anchor": 65, "temporal_embedding": 66, "third_person_reference": 67, "tricolon": 68, "us_them": 69, "vocative": 70 }, "layer_norm_eps": 1e-05, "layer_types": [ "full_attention", "sliding_attention", "sliding_attention", "full_attention", "sliding_attention", "sliding_attention", "full_attention", "sliding_attention", "sliding_attention", "full_attention", "sliding_attention", "sliding_attention", "full_attention", "sliding_attention", "sliding_attention", "full_attention", "sliding_attention", "sliding_attention", "full_attention", "sliding_attention", "sliding_attention", "full_attention" ], "local_attention": 128, "max_position_embeddings": 8192, "mlp_bias": false, "mlp_dropout": 0.0, "model_type": "modernbert", "norm_bias": false, "norm_eps": 1e-05, "num_attention_heads": 12, "num_hidden_layers": 22, "pad_token_id": 50283, "position_embedding_type": "absolute", "repad_logits_with_grad": false, "rope_parameters": { "full_attention": { "rope_theta": 160000.0, "rope_type": "default" }, "sliding_attention": { "rope_theta": 10000.0, "rope_type": "default" } }, "sep_token_id": 50282, "sparse_pred_ignore_index": -100, "sparse_prediction": false, "tie_word_embeddings": true, "transformers_version": "5.0.0", "vocab_size": 50368 }