electra.base.yaml

# args for data_transfer.py
data_path: "data/"
dev_path: "IWSLT12.TALK.dev2010.en-fr.en.xml" 
output_dev_path: 'dev_texts.txt'

train_path: "train.tags.en-fr.en.xml"
output_train_path: 'train_texts.txt'

test_path: "IWSLT12.TED.MT.tst2012.en-fr.en.xml"
output_test_path: 'test_texts_2012.txt'

# args for data_preprocess.py
data_path: "data/"
output_train_tsv: 'train.tsv'
output_test_tsv: 'test.tsv'
output_dev_tsv: 'dev.tsv'

# model hyper-parameters
batch_size: 128
device: 'gpu'
num_train_epochs: 1 
warmup_steps: 0
max_seq_length: 128
model_name_or_path: 'electra-base'
max_steps: -1
learning_rate: 5e-5
adam_epsilon: 1e-8
weight_decay: 0.0

# training parameters
global_step :  0
logging_steps: 200 # 日志的保存周期
save_steps: 200 # 模型保存周期
output_dir: 'checkpoints/' # 模型保存目录
ignore_label: -100

# args for predict.py
best_model: 'model_1000.pdparams' # 加载训练好的模型
output_pred_path: "results.txt"
isSavingPreds: 1 # 1: True