data:
  dataset_type: Ernie
  train_path: data/iwslt2012_zh/train.txt
  dev_path: data/iwslt2012_zh/dev.txt
  test_path: data/iwslt2012_zh/test.txt
  data_params: 
    pretrained_token: ernie-1.0
    punc_path: data/iwslt2012_zh/punc_vocab
    seq_len: 100
  batch_size: 64
  sortagrad: True 
  shuffle_method: batch_shuffle
  num_workers: 0

checkpoint:
  kbest_n: 5
  latest_n: 10
  metric_type: F1

model_type: ErnieLinear

model_params:
  pretrained_token: ernie-1.0
  num_classes: 4

training:
  n_epoch: 100
  lr: !!float 1e-5
  lr_decay: 1.0 
  weight_decay: !!float 1e-06
  global_grad_clip: 5.0
  log_interval: 10
  log_path: log/train_ernie_linear.log

testing:
  log_path: log/test_ernie_linear.log