search

DC-research / TEMPO

The official code for "TEMPO: Prompt-based Generative Pre-trained Transformer for Time Series Forecasting (ICLR 2024)". TEMPO is one of the very first open source Time Series Foundation Models for forecasting task v1.0 version.
78 stars 12 forks source link

Missing eval_data Attribute in Argument Parser – Error When Running etth1.sh Script #8

Closed cccccrj closed 6 days ago

cccccrj commented 2 weeks ago

When running the etth1.sh script in your scripts folder, I encountered the following error:

Namespace(batch_size=256, c_out=1, checkpoints='./lora_revin_6domain_checkpoints_1/', config_path='./configs/multiple_datasets.yml', cos=1, d_ff=768, d_model=768, datasets='ETTm1,ETTh2,ETTm2,electricity,traffic,weather', decay_fac=0.5, dropout=0.3, e_layers=3, electri_multiplier=1, embed='timeF', enc_in=7, equal=1, freeze=1, gpt_layers=6, hid_dim=16, is_gpt=1, itr=1, kernel_size=25, label_len=168, learning_rate=0.001, loss_func='mse', lradj='type3', max_len=-1, model='TEMPO', model_id='etth1_TEMPO_6_prompt_learn_336_96_100', n_heads=4, no_stl_loss=False, num_nodes=1, num_workers=0, patch_size=16, patience=5, pool=False, pred_len=96, pretrain=1, prompt=1, seq_len=336, stl_weight=0.001, stride=8, target_data='ETTh1', task_name='long_term_forecast', tmax=20, traffic_multiplier=1, train_epochs=10, use_token=0)
Traceback (most recent call last):
  File "main_multi_6domain_release.py", line 252, in <module>
    train_data, train_loader, test_data, test_loader, vali_data, vali_loader = prepare_data_loaders(args, config)
  File "main_multi_6domain_release.py", line 88, in prepare_data_loaders
    for dataset_name in args.eval_data.split(','):  
AttributeError: 'Namespace' object has no attribute 'eval_data'

I then checked the argparse module's help information but could not find any reference to eval_data. Could you provide some guidance on how to resolve this issue?

idevede commented 2 weeks ago

Thanks for your interest!

The eval_data attribute has been restored in the latest code update. We have updated it! Please let us know if you encounter any other issues.

Best