Is there any walk-through case to train custom dataset?

[forhonourlx]

Hi,

Is there any walk-through case to train custom dataset? I get confused with some config error of training custom daily timeseries data, like:

python -u run.py  --model_id DS_IC_test --is_training 1 --task_name in_context_forecast --root_path ./dataset/DS_IC --data_path DS_IC.csv --test_data_path DS_IC.csv --data IC --model AutoTimes_Llama --llm_ckp_dir E:\LLM\xyh110\Llama3-Chinese-8B-256K --gpu 0 --num_workers 20 --batch_size 32 --train_epochs 10 --mlp_hidden_dim 512 --learning_rate 0.0001 --weight_decay 0.00001 --loss SMAPE --use_amp --cosine --tmax 10 --seasonal_patterns Daily --seq_len 1024 --label_len 512 --test_pred_len 24 

Thanks inadvance. Anyway, if you are very busy, then let's look forward to the next update.

[WenWeiTHU]

Hello, could you please provide more error information to assist in resolving the issue?