URL-TITLE-classifier / config.json
tshasan's picture
Upload folder using huggingface_hub
e0f7eb6 verified
raw
history blame
1.88 kB
{
"_attn_implementation_autoset": true,
"_name_or_path": "firefoxrecap/URL-TITLE-classifier",
"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": 768,
"id2label": {
"0": "News",
"1": "Entertainment",
"2": "Shop",
"3": "Chat",
"4": "Education",
"5": "Government",
"6": "Health",
"7": "Technology",
"8": "Work",
"9": "Travel",
"10": "Uncategorized"
},
"initializer_cutoff_factor": 2.0,
"initializer_range": 0.02,
"intermediate_size": 1152,
"label2id": {
"Chat": 3,
"Education": 4,
"Entertainment": 1,
"Government": 5,
"Health": 6,
"News": 0,
"Shop": 2,
"Technology": 7,
"Travel": 9,
"Uncategorized": 10,
"Work": 8
},
"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": 12,
"num_hidden_layers": 22,
"pad_token_id": 50283,
"position_embedding_type": "absolute",
"problem_type": "multi_label_classification",
"reference_compile": true,
"repad_logits_with_grad": false,
"sep_token_id": 50282,
"sparse_pred_ignore_index": -100,
"sparse_prediction": false,
"torch_dtype": "float32",
"transformers_version": "4.49.0",
"vocab_size": 50368
}