map_c / config.json
qwertyuiopasdfg's picture
Upload folder using huggingface_hub
6f46b9d verified
{
"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,
"embedding_dropout": 0.0,
"eos_token_id": 50282,
"global_attn_every_n_layers": 3,
"global_rope_theta": 160000.0,
"gradient_checkpointing": false,
"hidden_activation": "gelu",
"hidden_size": 1024,
"id2label": {
"0": "Adding_across",
"1": "Adding_terms",
"2": "Additive",
"3": "Base_rate",
"4": "Certainty",
"5": "Correct",
"6": "Definition",
"7": "Denominator-only_change",
"8": "Division",
"9": "Duplication",
"10": "Firstterm",
"11": "FlipChange",
"12": "Ignores_zeroes",
"13": "Incomplete",
"14": "Incorrect_equivalent_fraction_addition",
"15": "Interior",
"16": "Inverse_operation",
"17": "Inversion",
"18": "Irrelevant",
"19": "Longer_is_bigger",
"20": "Mult",
"21": "Multiplying_by_4",
"22": "Neither",
"23": "Not_variable",
"24": "Positive",
"25": "Scale",
"26": "Shorter_is_bigger",
"27": "Subtraction",
"28": "SwapDividend",
"29": "Tacking",
"30": "Unknowable",
"31": "WNB",
"32": "Whole_numbers_larger",
"33": "Wrong_Fraction",
"34": "Wrong_Operation",
"35": "Wrong_fraction",
"36": "Wrong_term"
},
"initializer_cutoff_factor": 2.0,
"initializer_range": 0.02,
"intermediate_size": 2624,
"label2id": {
"Adding_across": 0,
"Adding_terms": 1,
"Additive": 2,
"Base_rate": 3,
"Certainty": 4,
"Correct": 5,
"Definition": 6,
"Denominator-only_change": 7,
"Division": 8,
"Duplication": 9,
"Firstterm": 10,
"FlipChange": 11,
"Ignores_zeroes": 12,
"Incomplete": 13,
"Incorrect_equivalent_fraction_addition": 14,
"Interior": 15,
"Inverse_operation": 16,
"Inversion": 17,
"Irrelevant": 18,
"Longer_is_bigger": 19,
"Mult": 20,
"Multiplying_by_4": 21,
"Neither": 22,
"Not_variable": 23,
"Positive": 24,
"Scale": 25,
"Shorter_is_bigger": 26,
"Subtraction": 27,
"SwapDividend": 28,
"Tacking": 29,
"Unknowable": 30,
"WNB": 31,
"Whole_numbers_larger": 32,
"Wrong_Fraction": 33,
"Wrong_Operation": 34,
"Wrong_fraction": 35,
"Wrong_term": 36
},
"layer_norm_eps": 1e-05,
"local_attention": 128,
"local_rope_theta": 10000.0,
"max_position_embeddings": 8192,
"mlp_bias": false,
"mlp_dropout": 0.0,
"model_type": "modernbert",
"norm_bias": false,
"norm_eps": 1e-05,
"num_attention_heads": 16,
"num_hidden_layers": 28,
"pad_token_id": 50283,
"position_embedding_type": "absolute",
"problem_type": "single_label_classification",
"repad_logits_with_grad": false,
"sep_token_id": 50282,
"sparse_pred_ignore_index": -100,
"sparse_prediction": false,
"torch_dtype": "bfloat16",
"transformers_version": "4.54.1",
"vocab_size": 50368
}