There is an issue with reading the load file

[fsmss]

Why is there an error in the parameter handling of this code?

[qiu69]

Thank you for your question. I'm sorry I didn't fully understand your meaning. Would it be convenient for you to take a screenshot.

[YEKAI-2022]

Hello, may I ask why this error is being reported?

[qiu69]

@YEKAI-2022 Can you tell me your running script？

[YEKAI-2022]

@YEKAI-2022 Can you tell me your running script？
在pycharm中的命令
--config-path
"rolling_forecast_config.json"
--data-name-list
"dataset/ETT-small/ETTh1.csv"
--strategy-args
'{"horizon":96}'
--model-name
"duet.DUET"
--model-hyper-params
'{"batch_size":
32,
"d_ff":
512,
"d_model":
512,
"dropout":
0.5,
"fc_dropout":
0.1,"n_heads":
1,
"e_layers":
1,
"lr":
0.0005,
"horizon":
96,
"seq_len":
512,
"factor":
3,
"lradj":
"type1",
"loss":
"MAE",
"num_experts":
2,
"k":
1,
"patch_len":
48,
"patience":
5,
"num_epochs":
100,
"CI":
1}'
--gpus
0
--num-workers
0
--timeout
60000
--save-path
"ETTh1/DUET"

[fsmss]

@YEKAI-2022 Can you tell me your running script？

I made the same mistake as him, the script seems to have used the wrong JSON format, using the script you provided.

[qiu69]

please refer to: https://github.com/decisionintelligence/DUET, FAQ: Q1

[YEKAI-2022]

我下好了数据集为什么路径错了，--data-name-list"E:\AI\Time series analysis\DUET-main\DUET-main\dataset\ETT-small\ETTh1.csv"

[YEKAI-2022]

请问下载好的数据集的路径该怎么指定呀

[fsmss]

请问下载好的数据集的路径该怎么指定呀
相同的问题 好像没有把路径传过去

[qiu69]

您好，请在DUET的readme里面下载DUET的数据集。然后只要指定数据集的名称就行，不用写路径。

[qiu69]

[qiu69]

请问问题得到解决了嘛？

[fsmss]

请问问题得到解决了嘛？

能够训练 但是结束的时候报错了

[qiu69]

请问跑出结果了吗，以及您运行的脚本方便发一下吗？

[fsmss]

python ./scripts/run_benchmark.py --config-path "rolling_forecast_config.json" --data-name-list "ILI.csv" --strategy-args {"horizon":24} --model-name "duet.DUET" --model-hyper-params {"batch_size":8,"dropout":0.15,"fc_dropout":0,"d_ff":1024,"d_model":128,"n_heads":1,"e_layers":2,"lr":0.0005,"horizon":24,"seq_len":104,"factor":3,"lradj":"type1","loss":"MAE","num_experts":2,"k":2,"patch_len":48,"patience":5,"num_epochs":100,"CI":1} --gpus 0 --num-workers 1 --timeout 60000 --save-path "ILI/DUET"

生成了一个压缩包

[qiu69]

压缩包里面看mae_norm和mse_norm这两列就行，这也是论文中汇报的对应mae和mse结果，指在归一化后的数据集以及预测结果上面计算误差指标。压缩包里面的mae和mse对应是返归一化后计算误差指标。

[qiu69]

直接运行：sh ./scripts/multivariate_forecast/ETTh1_script/DUET.sh。看看是否报错。以及您方便加我微信吗，qxf780310.

[liukaiyueyuo]

@qiu69 请问：No such file or directory: './dataset/forecasting/FORECAST_META.csv'。这个FORECAST_META.csv是啥文件？哪里下载？ 为啥报错？

[qiu69]

@liukaiyueyuo 您好，readme里面有下载数据集的链接，FORECAST_META.csv这个文件已经被包含在内了。

[liruyungitplayer]

老师您好，我在pycharm终端运行脚本时，参数也报错了，并且找了很久也找不到原因，老师您能帮我看一下吗

[liruyungitplayer]

老师您好，我在pycharm终端运行脚本时，参数也报错了，并且找了很久也找不到原因，老师您能帮我看一下吗

已经改好了，在大括号外边加上单引号就可以了