export CUDA_VISIBLE_DEVICES=0 model_name=TimesNet use_ps_loss=$1 # Use MSE loss only 1: Use PS loss patch_len_threshold=24 if [ ! -d "./logs/MSE/" ];then mkdir -p ./logs/MSE/ fi if [ ! -d "./logs/PS/" ];then mkdir -p ./logs/PS/ fi if [ "$use_ps_loss" -eq 0 ]; then ps_lambdas=(0.0) loss_name=MSE else ps_lambdas=(1.0 3.0 5.0 10.0) loss_name=PS fi for ps_lambda in ${ps_lambdas[@]}; do python -u run.py \ --task_name long_term_forecast \ --is_training 1 \ --root_path ../datasets/electricity/ \ --data_path electricity.csv \ --model_id ECL_96_${pred_len} \ --model $model_name \ --data custom \ --features M \ --seq_len 96 \ --label_len 48 \ --pred_len ${pred_len} \ --e_layers 2 \ --d_layers 1 \ --factor 3 \ --enc_in 321 \ --dec_in 321 \ --c_out 321 \ --d_model 256 \ --d_ff 512 \ --top_k 5 \ --des 'Exp' \ --use_ps_loss $use_ps_loss\ --ps_lambda $ps_lambda\ --patch_len_threshold $patch_len_threshold\ --iter 1 >logs/${loss_name}/ECL_96_96_${ps_lambda}lambda.log python -u run.py \ --task_name long_term_forecast \ --is_training 1 \ --root_path ../datasets/electricity/ \ --data_path electricity.csv \ --model_id ECL_96_192 \ --model $model_name \ --data custom \ --features M \ --seq_len 96 \ --label_len 48 \ --pred_len 192 \ --e_layers 2 \ --d_layers 1 \ --factor 3 \ --enc_in 321 \ --dec_in 321 \ --c_out 321 \ --d_model 256 \ --d_ff 512 \ --top_k 5 \ --des 'Exp' \ --use_ps_loss $use_ps_loss\ --ps_lambda $ps_lambda\ --patch_len_threshold $patch_len_threshold\ --iter 1 >logs/${loss_name}/ECL_96_192_${ps_lambda}lambda.log python -u run.py \ --task_name long_term_forecast \ --is_training 1 \ --root_path ../datasets/electricity/ \ --data_path electricity.csv \ --model_id ECL_96_336 \ --model $model_name \ --data custom \ --features M \ --seq_len 96 \ --label_len 48 \ --pred_len 336 \ --e_layers 2 \ --d_layers 1 \ --factor 3 \ --enc_in 321 \ --dec_in 321 \ --c_out 321 \ --d_model 256 \ --d_ff 512 \ --top_k 5 \ --des 'Exp' \ --use_ps_loss $use_ps_loss\ --ps_lambda $ps_lambda\ --patch_len_threshold $patch_len_threshold\ --iter 1 >logs/${loss_name}/ECL_96_336_${ps_lambda}lambda.log python -u run.py \ --task_name long_term_forecast \ --is_training 1 \ --root_path ../datasets/electricity/ \ --data_path electricity.csv \ --model_id ECL_96_720 \ --model $model_name \ --data custom \ --features M \ --seq_len 96 \ --label_len 48 \ --pred_len 720 \ --e_layers 2 \ --d_layers 1 \ --factor 3 \ --enc_in 321 \ --dec_in 321 \ --c_out 321 \ --d_model 256 \ --d_ff 512 \ --top_k 5 \ --des 'Exp' \ --use_ps_loss $use_ps_loss\ --ps_lambda $ps_lambda\ --patch_len_threshold $patch_len_threshold\ --iter 1 >logs/${loss_name}/ECL_96_720_${ps_lambda}lambda.log done