import torch

class Config:
    def __init__(self):
        # 1- 设备
        # self.device = ("cuda" if torch.cuda.is_available() else "cpu")
        # 下面的代码为了DQ量化
        self.device = "cpu"

        # 2- 原始文件路径
        self.train_datapath = "../00-data/data/train.txt"
        self.dev_datapath = "../00-data/data/dev.txt"
        self.test_datapath = "../00-data/data/test.txt"
        self.class_datapath = "../00-data/data/class.txt"

        # 3- 数据加载器参数
        self.batch_size = 64    # 64条新闻
        self.max_length = 32    # 句子中词的个数最多是32个词

        # 4- Bert预训练模型的路径
        # self.bert_path = r"D:\code\workspace7\ky5_NLP\models\bert-base-chinese"
        self.bert_path = r"D:\code\workspace8\TMF\03-bert\bert-base-chinese"

        # 5- 目标值文件解析
        self.classname_list = [line.strip() for line in open(self.class_datapath,mode="r",encoding="UTF-8")]
        self.classname_len = len(self.classname_list)

        # 6- 日志文件存储目录
        self.log_path = "./"

        # 7- 训练好模型的保存路径
        # .a  .abc  .xxx 障眼法
        self.save_model = "save_model/bert.pkl"


if __name__ == '__main__':
    config = Config()
    print(config.classname_list)
    print(config.classname_len)
    print(config.bert_path)