gliner_config.json

{
  "class_token_index": 128002,
  "dropout": 0.1,
  "encoder_config": null,
  "ent_token": "<<ENT>>",
  "eval_every": 1000,
  "fine_tune": true,
  "freeze_token_rep": false,
  "has_rnn": true,
  "hidden_size": 512,
  "log_dir": "logs_final",
  "loss_alpha": 0.75,
  "loss_gamma": 0,
  "loss_reduction": "sum",
  "lr_encoder": "5e-6",
  "lr_others": "7e-6",
  "max_len": 768,
  "max_neg_type_ratio": 1,
  "max_types": 30,
  "max_width": 100,
  "model_name": "microsoft/deberta-v3-large",
  "model_type": "gliner",
  "name": "token level gliner large",
  "num_steps": 6000,
  "prev_path": "logs_large/model_120000",
  "random_drop": true,
  "root_dir": "gliner_logs",
  "save_total_limit": 10,
  "scheduler_type": "linear",
  "sep_token": "<<SEP>>",
  "shuffle_types": true,
  "size_sup": -1,
  "span_mode": "token_level",
  "subtoken_pooling": "first",
  "train_batch_size": 8,
  "train_data": "../../data/unie_ner1.json",
  "transformers_version": "4.38.2",
  "val_data_dir": "none",
  "vocab_size": 128004,
  "warmup_ratio": 0.1,
  "weight_decay_encoder": 0.01,
  "weight_decay_other": 0.01,
  "words_splitter_type": "whitespace"
}