| { |
| "CUDA_VISIBLE_DEVICES": "0", |
| "USE_TORCH": "1", |
| "output_hidden_states": null, |
| "pretrained_model_name_or_path": "hfl/chinese-macbert-base", |
| "model_save_path": "../output/text_correction/model_public_csc", |
| "config_name": "csc.config", |
| "model_name": "pytorch_model.bin", |
| "path_train": "csc_public.train.json", |
| "path_dev": "csc_public.dev.json", |
| "path_tet": "csc_public.tet.json", |
| "scheduler_name": "cosine", |
| "tokenizer_type": "CHAR", |
| "padding_side": "RIGHT", |
| "active_type": "RELU", |
| "task_type": "CSC", |
| "model_type": "BERT", |
| "loss_type": "BCE", |
| "loss_det_rate": 0.3, |
| "max_len_limit": 512, |
| "batch_size": 32, |
| "num_labels": 0, |
| "max_len": 128, |
| "epochs": 3, |
| "lr": 3e-05, |
| "grad_accum_steps": 4, |
| "max_grad_norm": 1.0, |
| "weight_decay": 0.01, |
| "dropout_rate": 0.1, |
| "adam_eps": 1e-08, |
| "seed": 42, |
| "evaluate_steps": 1000, |
| "warmup_steps": 0.1, |
| "ignore_index": 0, |
| "save_steps": 1000, |
| "stop_epochs": 4, |
| "num_workers": 0, |
| "max_steps": -1, |
| "flag_save_model_state": true, |
| "flag_dynamic_encode": false, |
| "flag_tokenizer_char": true, |
| "flag_soft_label": true, |
| "flag_save_best": true, |
| "flag_dropout": false, |
| "flag_shuffle": true, |
| "flag_active": false, |
| "flag_train": false, |
| "flag_cuda": true, |
| "flag_mft": true, |
| "flag_adv": false, |
| "xy_keys_predict": [ |
| "original_text", |
| "correct_text", |
| "wrong_ids" |
| ], |
| "keys": [ |
| "original_text", |
| "correct_text", |
| "wrong_ids" |
| ], |
| "save_best_mertics_key": [ |
| "sentence", |
| "strict_cor_f1" |
| ], |
| "label_sep": "|myz|", |
| "multi_label_threshold": 0.5, |
| "len_rate": 1, |
| "adv_emb_name": "word_embeddings.", |
| "adv_eps": 1.0, |
| "additional_special_tokens": [], |
| "len_corpus": null, |
| "prior_count": null, |
| "prior": null, |
| "l2i": null, |
| "i2l": null, |
| "xy_keys": [ |
| "original_text", |
| "correct_text", |
| "wrong_ids" |
| ] |
| } |
