update model

mirror_outputs/Mirror_Pretrain_AllExcluded_2/ckpt/SchemaGuidedInstructBertModel.best.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5e4e49b84688218d2e60b1123682e313478cd796739dbe61a035890600e324c0
+size 5229689445

mirror_outputs/Mirror_Pretrain_AllExcluded_2/task_params.yaml

@@ -0,0 +1,56 @@
+_config_info:
+  create_time: '2023-06-16 12:37:14'
+  use_default_base_config: true
+  config_filepath:
+  - /data/tzhu/Mirror/conf/Pretrain_excluded.yaml
+task_type: SchemaGuidedInstructBertTask
+task_name: Mirror_Pretrain_AllExcluded_2
+comment: ~~content as label, (start, end + 1) span
+output_dir: mirror_outputs
+task_dir: mirror_outputs/Mirror_Pretrain_AllExcluded_2
+dump_cache_dir: mirror_outputs/Mirror_Pretrain_AllExcluded_2/cache
+regenerate_cache: false
+data_dir: resources/Mirror/v1.4_sampled_v3/merged/all_excluded
+train_filepath: resources/Mirror/v1.4_sampled_v3/merged/all_excluded/train.jsonl
+dev_filepath: resources/Mirror/v1.4_sampled_v3/merged/all_excluded/dev.jsonl
+test_filepath: resources/Mirror/v1.4_sampled_v3/merged/all_excluded/test.jsonl
+random_seed: 1227
+num_epochs: 3
+num_steps: -1
+warmup_proportion: 0.1
+epoch_patience: -1
+step_patience: -1
+batch_size: 64
+learning_rate: 2.0e-05
+max_grad_norm: 1.0
+skip_train: false
+debug_mode: false
+grad_accum_steps: 1
+resumed_training_path: null
+step_eval_interval: 10000
+epoch_eval_interval: 1
+eval_on_data:
+- train
+select_best_on_data: train
+select_best_by_key: loss
+best_metric_field: micro.f1
+save_every_ckpt: true
+save_best_ckpt: true
+final_eval_on_test: false
+main_process_logging: true
+max_seq_len: 512
+label_span: tag
+mode: span
+stream_mode: false
+plm_dir: /data/tzhu/PLM/microsoft--deberta-v3-large
+base_model_path: null
+train_batch_size: 8
+eval_batch_size: 8
+other_learning_rate: 0.0001
+weight_decay: 0.1
+dropout: 0.3
+use_rope: true
+biaffine_size: 512
+include_package:
+- src.task
+use_default_base_config: true