RISys-Lab/CyberSec-Text-Classification-ModernBert-Base
064
1{2 "_name_or_path": "answerdotai/ModernBERT-base",3 "_num_labels": 2,4 "architectures": [5 "ModernBertForSequenceClassification"6 ],7 "attention_bias": false,8 "attention_dropout": 0.0,9 "bos_token_id": 50281,10 "classifier_activation": "gelu",11 "classifier_bias": false,12 "classifier_dropout": 0.0,13 "classifier_pooling": "mean",14 "cls_token_id": 50281,15 "decoder_bias": true,16 "deterministic_flash_attn": false,17 "embedding_dropout": 0.0,18 "eos_token_id": 50282,19 "global_attn_every_n_layers": 3,20 "global_rope_theta": 160000.0,21 "gradient_checkpointing": false,22 "hidden_activation": "gelu",23 "hidden_size": 768,24 "id2label": {25 "0": "non-cybersec",26 "1": "cybersec"27 },28 "initializer_cutoff_factor": 2.0,29 "initializer_range": 0.02,30 "intermediate_size": 1152,31 "label2id": {32 "cybersec": 1,33 "non-cybersec": 034 },35 "layer_norm_eps": 1e-05,36 "local_attention": 128,37 "local_rope_theta": 10000.0,38 "max_position_embeddings": 8192,39 "mlp_bias": false,40 "mlp_dropout": 0.0,41 "model_type": "modernbert",42 "norm_bias": false,43 "norm_eps": 1e-05,44 "num_attention_heads": 12,45 "num_hidden_layers": 22,46 "pad_token_id": 50283,47 "position_embedding_type": "absolute",48 "problem_type": "single_label_classification",49 "reference_compile": true,50 "sep_token_id": 50282,51 "sparse_pred_ignore_index": -100,52 "sparse_prediction": false,53 "torch_dtype": "float32",54 "transformers_version": "4.48.0.dev0",55 "vocab_size": 5036856}57 