Skip to content

[🐞] 关于使用DLinear模型评估得到的MAE,RMSE和MAPE指标 #298

@wubo2180

Description

@wubo2180

Is there an existing issue / discussion for this? | 是否已有关于该错误的issue或讨论?

  • I have searched the existing issues / discussions | 我已经搜索过已有的issues和讨论

Is there an existing answer for this in tutorial? | 该问题是否在教程中有解答?

  • I have searched tutorial | 我已经搜索过tutorial

Current Behavior | 当前行为

您好,感谢您提供的框架代码,我使用models中的DLinear模型在跑METR-LA数据集时,发现训练得到MAE/RMSE/MAPE都远低于一些Baselines的值

Expected Behavior | 期望行为

Image

No response

Environment | 运行环境

- OS:Ubuntu 22.04.2
- DEVICE:NVIDIA GeForce RTX 5090
- NVIDIA Driver:580.76.05
- CUDA:13.0
- NVIDIA GPU Memory: 32GB
- PyTorch:2.8.0+cu128

BasicTS logs | BasicTS日志

Image

Steps To Reproduce | 复现方法

import os
from pathlib import Path
current_file = Path(__file__).resolve()
src_dir = current_file.parent.parent
if str(src_dir) not in sys.path:
    sys.path.insert(0, str(src_dir))
project_root = src_dir.parent 
os.chdir(project_root)
from basicts.models.DLinear import DLinear, DLinearConfig
from basicts.configs import BasicTSForecastingConfig
from basicts import BasicTSLauncher
def main():
    dataset_name = 'METR-LA'
    model_config = DLinearConfig(input_len=12, 
                                 output_len=12)
    cfg = BasicTSForecastingConfig(
        model=DLinear,
        model_config = model_config,
        dataset_name = dataset_name,
        gpus = "0",
        input_len = 12,
        output_len = 12,
        batch_size = 64,
        use_timestamps = True,
    )
    BasicTSLauncher.launch_training(cfg)
if __name__ == "__main__":
    main()

Anything else? | 备注

No response

Metadata

Metadata

Assignees

No one assigned

    Labels

    bugSomething isn't working

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions