Closed hughxx closed 1 year ago
ssh://hugh@202.117.21.53:22/home/hugh/anaconda3/bin/python -u /home/share/hugh/project/timesnet/run_itrans.py Args in experiment: Namespace(activation='gelu', anomaly_ratio=0.25, batch_size=16, c_out=321, checkpoints='./checkpoints/', d_ff=512, d_layers=1, d_model=512, data='custom', data_path='weather.csv', dec_in=321, des='test', devices='0,1,2,3', distil=True, dropout=0.1, e_layers=3, embed='timeF', enc_in=321, factor=3, features='M', freq='h', gpu=0, inverse=False, is_training=1, itr=1, label_len=48, learning_rate=0.0005, loss='MSE', lradj='type1', mask_rate=0.25, model='iTransformer', model_id='ECL_96', moving_avg=25, n_heads=8, num_kernels=6, num_workers=10, output_attention=False, p_hidden_dims=[128, 128], p_hidden_layers=2, patience=3, pred_len=96, root_path='./dataset/', seasonal_patterns='Monthly', seq_len=96, target='OT', task_name='long_term_forecast', top_k=5, train_epochs=10, use_amp=False, use_gpu=True, use_multi_gpu=False) Use GPU: cuda:0 Number of parameters: 4.83M
start training : long_term_forecast_ECL_96_iTransformer_custom_ftM_sl96_ll48_pl96_dm512_nh8_el3_dl1_df512_fc3_ebtimeF_dtTrue_test_0>>>>>>>>>>>>>>>>>>>>>>>>>> train 36696 val 5175 test 10444 iters: 100, epoch: 1 | loss: 0.4884371 speed: 0.0937s/iter; left time: 2139.7748s iters: 200, epoch: 1 | loss: 0.3802813 speed: 0.0608s/iter; left time: 1383.0623s iters: 300, epoch: 1 | loss: 0.4912418 speed: 0.0675s/iter; left time: 1527.0099s iters: 400, epoch: 1 | loss: 0.6137630 speed: 0.0755s/iter; left time: 1701.1963s iters: 500, epoch: 1 | loss: 0.3286369 speed: 0.0653s/iter; left time: 1464.2194s iters: 600, epoch: 1 | loss: 0.5121265 speed: 0.0600s/iter; left time: 1339.2188s iters: 700, epoch: 1 | loss: 0.3868058 speed: 0.0622s/iter; left time: 1383.5268s iters: 800, epoch: 1 | loss: 0.4105329 speed: 0.0653s/iter; left time: 1446.1897s iters: 900, epoch: 1 | loss: 0.8757164 speed: 0.0598s/iter; left time: 1317.4781s iters: 1000, epoch: 1 | loss: 0.6362411 speed: 0.0591s/iter; left time: 1295.4367s iters: 1100, epoch: 1 | loss: 0.8004262 speed: 0.0602s/iter; left time: 1314.6760s iters: 1200, epoch: 1 | loss: 0.4904622 speed: 0.0629s/iter; left time: 1367.5994s iters: 1300, epoch: 1 | loss: 0.8895156 speed: 0.0614s/iter; left time: 1328.3168s iters: 1400, epoch: 1 | loss: 0.3681978 speed: 0.0567s/iter; left time: 1220.0680s iters: 1500, epoch: 1 | loss: 0.6627781 speed: 0.0534s/iter; left time: 1143.8895s iters: 1600, epoch: 1 | loss: 0.4666792 speed: 0.0576s/iter; left time: 1228.0221s iters: 1700, epoch: 1 | loss: 0.5510955 speed: 0.0634s/iter; left time: 1345.7580s iters: 1800, epoch: 1 | loss: 0.6077036 speed: 0.0631s/iter; left time: 1333.4629s iters: 1900, epoch: 1 | loss: 0.7905897 speed: 0.0618s/iter; left time: 1300.6507s iters: 2000, epoch: 1 | loss: 0.4359589 speed: 0.0644s/iter; left time: 1348.0659s iters: 2100, epoch: 1 | loss: 0.4578832 speed: 0.0589s/iter; left time: 1226.9304s iters: 2200, epoch: 1 | loss: 0.3511912 speed: 0.0648s/iter; left time: 1342.4648s Epoch: 1 cost time: 145.5831642150879 Epoch: 1, Steps: 2293 | Train Loss: 0.6292579 Vali Loss: 0.6010199 Test Loss: 0.2191115 Validation loss decreased (inf --> 0.601020). Saving model ... Updating learning rate to 0.0005 iters: 100, epoch: 2 | loss: 0.4294548 speed: 0.9144s/iter; left time: 18779.3912s iters: 200, epoch: 2 | loss: 0.4692017 speed: 0.0769s/iter; left time: 1570.8470s iters: 300, epoch: 2 | loss: 0.5572400 speed: 0.0745s/iter; left time: 1515.8194s iters: 400, epoch: 2 | loss: 0.5315792 speed: 0.0772s/iter; left time: 1561.5736s iters: 500, epoch: 2 | loss: 0.6228209 speed: 0.0777s/iter; left time: 1565.6174s iters: 600, epoch: 2 | loss: 0.4701141 speed: 0.0794s/iter; left time: 1590.5029s iters: 700, epoch: 2 | loss: 0.3707439 speed: 0.0725s/iter; left time: 1445.7180s iters: 800, epoch: 2 | loss: 0.6635921 speed: 0.0709s/iter; left time: 1406.6114s iters: 900, epoch: 2 | loss: 0.4851658 speed: 0.0722s/iter; left time: 1425.9514s iters: 1000, epoch: 2 | loss: 0.4823830 speed: 0.0739s/iter; left time: 1451.9987s iters: 1100, epoch: 2 | loss: 0.5683275 speed: 0.0715s/iter; left time: 1396.3928s iters: 1200, epoch: 2 | loss: 0.6960465 speed: 0.0802s/iter; left time: 1558.5522s iters: 1300, epoch: 2 | loss: 0.6440760 speed: 0.0719s/iter; left time: 1391.2820s iters: 1400, epoch: 2 | loss: 0.4316963 speed: 0.0749s/iter; left time: 1440.1087s iters: 1500, epoch: 2 | loss: 2.5233083 speed: 0.0733s/iter; left time: 1403.4639s iters: 1600, epoch: 2 | loss: 0.5071030 speed: 0.0689s/iter; left time: 1311.2473s iters: 1700, epoch: 2 | loss: 0.4185750 speed: 0.0683s/iter; left time: 1293.9478s iters: 1800, epoch: 2 | loss: 0.4477855 speed: 0.0719s/iter; left time: 1355.2583s iters: 1900, epoch: 2 | loss: 0.5800842 speed: 0.0718s/iter; left time: 1345.0957s iters: 2000, epoch: 2 | loss: 0.4868422 speed: 0.0722s/iter; left time: 1344.9053s iters: 2100, epoch: 2 | loss: 0.6417049 speed: 0.0759s/iter; left time: 1407.4534s iters: 2200, epoch: 2 | loss: 0.4854120 speed: 0.0728s/iter; left time: 1342.4686s Epoch: 2 cost time: 169.77924752235413 Epoch: 2, Steps: 2293 | Train Loss: 0.6592892 Vali Loss: 0.6239563 Test Loss: 0.2195540 EarlyStopping counter: 1 out of 3 Updating learning rate to 0.00025 iters: 100, epoch: 3 | loss: 0.4906038 speed: 1.0929s/iter; left time: 19939.3801s iters: 200, epoch: 3 | loss: 0.9143455 speed: 0.0745s/iter; left time: 1351.8604s iters: 300, epoch: 3 | loss: 0.4077953 speed: 0.0745s/iter; left time: 1343.5740s iters: 400, epoch: 3 | loss: 0.5590039 speed: 0.0746s/iter; left time: 1337.9778s iters: 500, epoch: 3 | loss: 2.3589504 speed: 0.0809s/iter; left time: 1443.0674s iters: 600, epoch: 3 | loss: 0.4440592 speed: 0.0682s/iter; left time: 1210.9154s iters: 700, epoch: 3 | loss: 0.3892558 speed: 0.0766s/iter; left time: 1351.2589s iters: 800, epoch: 3 | loss: 0.4404001 speed: 0.0731s/iter; left time: 1281.8376s iters: 900, epoch: 3 | loss: 0.5981345 speed: 0.0745s/iter; left time: 1300.4966s iters: 1000, epoch: 3 | loss: 0.7439384 speed: 0.0741s/iter; left time: 1285.2254s iters: 1100, epoch: 3 | loss: 0.4372765 speed: 0.0787s/iter; left time: 1357.9086s iters: 1200, epoch: 3 | loss: 0.9567031 speed: 0.0763s/iter; left time: 1307.9602s iters: 1300, epoch: 3 | loss: 0.4343786 speed: 0.0737s/iter; left time: 1256.4248s iters: 1400, epoch: 3 | loss: 0.5687712 speed: 0.0797s/iter; left time: 1350.6784s iters: 1500, epoch: 3 | loss: 0.4656307 speed: 0.0848s/iter; left time: 1427.8978s iters: 1600, epoch: 3 | loss: 2.6003740 speed: 0.0855s/iter; left time: 1432.3080s iters: 1700, epoch: 3 | loss: 0.9539509 speed: 0.0801s/iter; left time: 1334.0940s iters: 1800, epoch: 3 | loss: 0.4689850 speed: 0.0810s/iter; left time: 1340.7400s iters: 1900, epoch: 3 | loss: 0.7294191 speed: 0.0779s/iter; left time: 1281.3714s iters: 2000, epoch: 3 | loss: 0.9128451 speed: 0.0794s/iter; left time: 1297.8293s iters: 2100, epoch: 3 | loss: 0.5720030 speed: 0.0811s/iter; left time: 1317.2313s iters: 2200, epoch: 3 | loss: 0.6068072 speed: 0.0781s/iter; left time: 1260.1362s Epoch: 3 cost time: 178.00309491157532 Epoch: 3, Steps: 2293 | Train Loss: 0.6532407 Vali Loss: 0.5844204 Test Loss: 0.2118002 Validation loss decreased (0.601020 --> 0.584420). Saving model ... Updating learning rate to 0.000125 iters: 100, epoch: 4 | loss: 0.8171552 speed: 1.1715s/iter; left time: 18687.2540s iters: 200, epoch: 4 | loss: 0.4375284 speed: 0.0756s/iter; left time: 1198.4875s iters: 300, epoch: 4 | loss: 0.5282884 speed: 0.0717s/iter; left time: 1129.6914s iters: 400, epoch: 4 | loss: 0.4568482 speed: 0.0775s/iter; left time: 1212.4092s iters: 500, epoch: 4 | loss: 0.5553384 speed: 0.0782s/iter; left time: 1215.5875s iters: 600, epoch: 4 | loss: 0.3938108 speed: 0.0775s/iter; left time: 1197.6517s iters: 700, epoch: 4 | loss: 0.5376428 speed: 0.0805s/iter; left time: 1236.4909s iters: 800, epoch: 4 | loss: 0.5792876 speed: 0.0718s/iter; left time: 1094.7297s iters: 900, epoch: 4 | loss: 1.0841678 speed: 0.0748s/iter; left time: 1133.6099s iters: 1000, epoch: 4 | loss: 0.5882192 speed: 0.0733s/iter; left time: 1102.7289s iters: 1100, epoch: 4 | loss: 0.4645280 speed: 0.0733s/iter; left time: 1096.2290s iters: 1200, epoch: 4 | loss: 0.5107097 speed: 0.0738s/iter; left time: 1096.1474s iters: 1300, epoch: 4 | loss: 0.4206262 speed: 0.0713s/iter; left time: 1051.4406s iters: 1400, epoch: 4 | loss: 0.6745049 speed: 0.0756s/iter; left time: 1106.9956s iters: 1500, epoch: 4 | loss: 0.6672496 speed: 0.0770s/iter; left time: 1120.6009s iters: 1600, epoch: 4 | loss: 0.9733395 speed: 0.0695s/iter; left time: 1004.9363s iters: 1700, epoch: 4 | loss: 0.5707955 speed: 0.0721s/iter; left time: 1035.2023s iters: 1800, epoch: 4 | loss: 0.3647259 speed: 0.0740s/iter; left time: 1054.5688s iters: 1900, epoch: 4 | loss: 0.4864632 speed: 0.0766s/iter; left time: 1083.8635s iters: 2000, epoch: 4 | loss: 0.6721052 speed: 0.0675s/iter; left time: 947.9728s iters: 2100, epoch: 4 | loss: 0.5320817 speed: 0.0812s/iter; left time: 1133.0953s iters: 2200, epoch: 4 | loss: 0.4686660 speed: 0.0735s/iter; left time: 1018.0475s Epoch: 4 cost time: 172.55122876167297 Epoch: 4, Steps: 2293 | Train Loss: 0.6197659 Vali Loss: 0.5653846 Test Loss: 0.2179509 Validation loss decreased (0.584420 --> 0.565385). Saving model ... Updating learning rate to 6.25e-05 iters: 100, epoch: 5 | loss: 1.3176355 speed: 1.0579s/iter; left time: 14450.4948s iters: 200, epoch: 5 | loss: 0.3632142 speed: 0.0783s/iter; left time: 1061.2547s iters: 300, epoch: 5 | loss: 0.5639267 speed: 0.0800s/iter; left time: 1077.3586s iters: 400, epoch: 5 | loss: 0.4628576 speed: 0.0802s/iter; left time: 1070.9687s iters: 500, epoch: 5 | loss: 0.3996843 speed: 0.0824s/iter; left time: 1092.5117s iters: 600, epoch: 5 | loss: 0.3751653 speed: 0.0867s/iter; left time: 1140.4472s iters: 700, epoch: 5 | loss: 0.4337548 speed: 0.0837s/iter; left time: 1093.0083s iters: 800, epoch: 5 | loss: 0.7628719 speed: 0.0841s/iter; left time: 1089.7295s iters: 900, epoch: 5 | loss: 0.3890971 speed: 0.0779s/iter; left time: 1001.7886s iters: 1000, epoch: 5 | loss: 0.4888734 speed: 0.0752s/iter; left time: 959.6902s iters: 1100, epoch: 5 | loss: 0.5164376 speed: 0.0663s/iter; left time: 839.1062s iters: 1200, epoch: 5 | loss: 0.7492502 speed: 0.0755s/iter; left time: 948.7595s iters: 1300, epoch: 5 | loss: 0.4169662 speed: 0.0776s/iter; left time: 966.5186s iters: 1400, epoch: 5 | loss: 0.5701704 speed: 0.0780s/iter; left time: 964.0657s iters: 1500, epoch: 5 | loss: 0.9501191 speed: 0.0781s/iter; left time: 956.8608s iters: 1600, epoch: 5 | loss: 0.3695846 speed: 0.0721s/iter; left time: 877.2684s iters: 1700, epoch: 5 | loss: 0.4854094 speed: 0.0770s/iter; left time: 928.2359s iters: 1800, epoch: 5 | loss: 0.4404168 speed: 0.0797s/iter; left time: 952.9556s iters: 1900, epoch: 5 | loss: 1.1658223 speed: 0.0830s/iter; left time: 984.8265s iters: 2000, epoch: 5 | loss: 0.9135388 speed: 0.0778s/iter; left time: 915.0434s iters: 2100, epoch: 5 | loss: 2.4204495 speed: 0.0707s/iter; left time: 824.0989s iters: 2200, epoch: 5 | loss: 0.7580358 speed: 0.0739s/iter; left time: 854.0770s Epoch: 5 cost time: 179.5488142967224 Epoch: 5, Steps: 2293 | Train Loss: 0.6026873 Vali Loss: 0.5452353 Test Loss: 0.2121598 Validation loss decreased (0.565385 --> 0.545235). Saving model ... Updating learning rate to 3.125e-05 iters: 100, epoch: 6 | loss: 0.5178975 speed: 1.0932s/iter; left time: 12424.9406s iters: 200, epoch: 6 | loss: 0.4163472 speed: 0.0849s/iter; left time: 956.2337s iters: 300, epoch: 6 | loss: 0.4349686 speed: 0.0839s/iter; left time: 936.3199s iters: 400, epoch: 6 | loss: 0.4697899 speed: 0.0783s/iter; left time: 866.2788s iters: 500, epoch: 6 | loss: 0.4088657 speed: 0.0826s/iter; left time: 905.2624s iters: 600, epoch: 6 | loss: 0.3523217 speed: 0.0832s/iter; left time: 903.9349s iters: 700, epoch: 6 | loss: 0.7175497 speed: 0.0831s/iter; left time: 894.2748s iters: 800, epoch: 6 | loss: 0.6137013 speed: 0.0820s/iter; left time: 874.2891s iters: 900, epoch: 6 | loss: 2.2971494 speed: 0.0887s/iter; left time: 936.8010s iters: 1000, epoch: 6 | loss: 0.2834386 speed: 0.0829s/iter; left time: 867.5681s iters: 1100, epoch: 6 | loss: 0.4658869 speed: 0.0809s/iter; left time: 839.1260s iters: 1200, epoch: 6 | loss: 0.6161052 speed: 0.0837s/iter; left time: 859.2096s iters: 1300, epoch: 6 | loss: 2.4163496 speed: 0.0794s/iter; left time: 807.6245s iters: 1400, epoch: 6 | loss: 0.3827536 speed: 0.0759s/iter; left time: 763.6435s iters: 1500, epoch: 6 | loss: 0.3719592 speed: 0.0797s/iter; left time: 793.8275s iters: 1600, epoch: 6 | loss: 0.5199808 speed: 0.0830s/iter; left time: 818.5331s iters: 1700, epoch: 6 | loss: 0.5045499 speed: 0.0795s/iter; left time: 776.4544s iters: 1800, epoch: 6 | loss: 0.3985899 speed: 0.0777s/iter; left time: 750.6160s iters: 1900, epoch: 6 | loss: 1.3146554 speed: 0.0836s/iter; left time: 799.8267s iters: 2000, epoch: 6 | loss: 0.8792403 speed: 0.0816s/iter; left time: 772.2885s iters: 2100, epoch: 6 | loss: 0.7732003 speed: 0.0799s/iter; left time: 748.3963s iters: 2200, epoch: 6 | loss: 0.3338433 speed: 0.0797s/iter; left time: 738.5635s Epoch: 6 cost time: 187.79173064231873 Epoch: 6, Steps: 2293 | Train Loss: 0.5963767 Vali Loss: 0.5462552 Test Loss: 0.2123922 EarlyStopping counter: 1 out of 3 Updating learning rate to 1.5625e-05 iters: 100, epoch: 7 | loss: 0.4431281 speed: 1.0908s/iter; left time: 9897.0948s iters: 200, epoch: 7 | loss: 0.3399965 speed: 0.0799s/iter; left time: 717.3346s iters: 300, epoch: 7 | loss: 0.3888144 speed: 0.0759s/iter; left time: 673.3893s iters: 400, epoch: 7 | loss: 0.4371489 speed: 0.0762s/iter; left time: 668.4013s iters: 500, epoch: 7 | loss: 0.4145016 speed: 0.0815s/iter; left time: 707.1688s iters: 600, epoch: 7 | loss: 0.4339801 speed: 0.0785s/iter; left time: 672.8625s iters: 700, epoch: 7 | loss: 0.3466100 speed: 0.0772s/iter; left time: 653.7720s iters: 800, epoch: 7 | loss: 0.3813485 speed: 0.0779s/iter; left time: 652.3099s iters: 900, epoch: 7 | loss: 0.3679409 speed: 0.0799s/iter; left time: 660.9565s iters: 1000, epoch: 7 | loss: 0.6518580 speed: 0.0779s/iter; left time: 636.5674s iters: 1100, epoch: 7 | loss: 0.7202061 speed: 0.0743s/iter; left time: 599.9129s iters: 1200, epoch: 7 | loss: 0.5429041 speed: 0.0777s/iter; left time: 619.7923s iters: 1300, epoch: 7 | loss: 0.6968318 speed: 0.0820s/iter; left time: 645.6435s iters: 1400, epoch: 7 | loss: 0.8312810 speed: 0.0783s/iter; left time: 608.5062s iters: 1500, epoch: 7 | loss: 0.4355338 speed: 0.0830s/iter; left time: 636.6275s iters: 1600, epoch: 7 | loss: 0.6479142 speed: 0.0780s/iter; left time: 590.5485s iters: 1700, epoch: 7 | loss: 0.3864921 speed: 0.0753s/iter; left time: 562.8512s iters: 1800, epoch: 7 | loss: 2.0052402 speed: 0.0763s/iter; left time: 562.5214s iters: 1900, epoch: 7 | loss: 0.3373819 speed: 0.0780s/iter; left time: 567.0853s iters: 2000, epoch: 7 | loss: 0.5641161 speed: 0.0784s/iter; left time: 562.6347s iters: 2100, epoch: 7 | loss: 0.4467024 speed: 0.0738s/iter; left time: 521.7921s iters: 2200, epoch: 7 | loss: 0.5322549 speed: 0.0811s/iter; left time: 565.6524s Epoch: 7 cost time: 180.17504382133484 Epoch: 7, Steps: 2293 | Train Loss: 0.5904231 Vali Loss: 0.5425926 Test Loss: 0.2129188 Validation loss decreased (0.545235 --> 0.542593). Saving model ... Updating learning rate to 7.8125e-06 iters: 100, epoch: 8 | loss: 0.5336107 speed: 1.0835s/iter; left time: 7346.1023s iters: 200, epoch: 8 | loss: 0.5202275 speed: 0.0783s/iter; left time: 523.1299s iters: 300, epoch: 8 | loss: 0.5076771 speed: 0.0774s/iter; left time: 509.0122s iters: 400, epoch: 8 | loss: 0.5141376 speed: 0.0746s/iter; left time: 483.5969s iters: 500, epoch: 8 | loss: 0.4696932 speed: 0.0742s/iter; left time: 473.3345s iters: 600, epoch: 8 | loss: 0.4274750 speed: 0.0797s/iter; left time: 500.4765s iters: 700, epoch: 8 | loss: 0.4773324 speed: 0.0765s/iter; left time: 472.8613s iters: 800, epoch: 8 | loss: 0.5086091 speed: 0.0754s/iter; left time: 458.3689s iters: 900, epoch: 8 | loss: 0.4554705 speed: 0.0684s/iter; left time: 409.0394s iters: 1000, epoch: 8 | loss: 0.3260530 speed: 0.0745s/iter; left time: 438.1234s iters: 1100, epoch: 8 | loss: 0.6249232 speed: 0.0754s/iter; left time: 435.6421s iters: 1200, epoch: 8 | loss: 0.4327333 speed: 0.0743s/iter; left time: 422.0205s iters: 1300, epoch: 8 | loss: 0.4534933 speed: 0.0760s/iter; left time: 423.9169s iters: 1400, epoch: 8 | loss: 0.9433058 speed: 0.0789s/iter; left time: 432.5085s iters: 1500, epoch: 8 | loss: 0.4841625 speed: 0.0787s/iter; left time: 423.5810s iters: 1600, epoch: 8 | loss: 0.5241430 speed: 0.0740s/iter; left time: 390.9475s iters: 1700, epoch: 8 | loss: 0.3432244 speed: 0.0721s/iter; left time: 373.5526s iters: 1800, epoch: 8 | loss: 0.5611855 speed: 0.0720s/iter; left time: 365.9353s iters: 1900, epoch: 8 | loss: 1.1085573 speed: 0.0790s/iter; left time: 393.6406s iters: 2000, epoch: 8 | loss: 2.0539229 speed: 0.0780s/iter; left time: 380.7078s iters: 2100, epoch: 8 | loss: 0.6593276 speed: 0.0783s/iter; left time: 374.1453s iters: 2200, epoch: 8 | loss: 0.6349206 speed: 0.0794s/iter; left time: 371.7328s Epoch: 8 cost time: 174.84419965744019 Epoch: 8, Steps: 2293 | Train Loss: 0.5879804 Vali Loss: 0.5418903 Test Loss: 0.2118427 Validation loss decreased (0.542593 --> 0.541890). Saving model ... Updating learning rate to 3.90625e-06 iters: 100, epoch: 9 | loss: 1.6399350 speed: 1.0972s/iter; left time: 4923.0557s iters: 200, epoch: 9 | loss: 0.4668166 speed: 0.0764s/iter; left time: 335.1979s iters: 300, epoch: 9 | loss: 2.3957839 speed: 0.0770s/iter; left time: 329.9856s iters: 400, epoch: 9 | loss: 0.7759449 speed: 0.0779s/iter; left time: 326.2490s iters: 500, epoch: 9 | loss: 0.3914853 speed: 0.0721s/iter; left time: 294.6746s iters: 600, epoch: 9 | loss: 0.3818814 speed: 0.0758s/iter; left time: 302.3805s iters: 700, epoch: 9 | loss: 0.8356951 speed: 0.0739s/iter; left time: 287.2303s iters: 800, epoch: 9 | loss: 0.3929401 speed: 0.0732s/iter; left time: 277.3072s iters: 900, epoch: 9 | loss: 0.6580318 speed: 0.0747s/iter; left time: 275.5474s iters: 1000, epoch: 9 | loss: 3.0244946 speed: 0.0752s/iter; left time: 269.5911s iters: 1100, epoch: 9 | loss: 0.4308555 speed: 0.0751s/iter; left time: 261.7653s iters: 1200, epoch: 9 | loss: 0.3591655 speed: 0.0729s/iter; left time: 246.8788s iters: 1300, epoch: 9 | loss: 0.3947488 speed: 0.0746s/iter; left time: 245.2144s iters: 1400, epoch: 9 | loss: 0.5670210 speed: 0.0717s/iter; left time: 228.4135s iters: 1500, epoch: 9 | loss: 0.6620324 speed: 0.0740s/iter; left time: 228.3872s iters: 1600, epoch: 9 | loss: 0.4885000 speed: 0.0681s/iter; left time: 203.5123s iters: 1700, epoch: 9 | loss: 0.4203905 speed: 0.0678s/iter; left time: 195.8633s iters: 1800, epoch: 9 | loss: 0.6348040 speed: 0.0692s/iter; left time: 192.7492s iters: 1900, epoch: 9 | loss: 0.5961450 speed: 0.0731s/iter; left time: 196.4392s iters: 2000, epoch: 9 | loss: 0.6816207 speed: 0.0725s/iter; left time: 187.6746s iters: 2100, epoch: 9 | loss: 0.5948603 speed: 0.0719s/iter; left time: 178.7357s iters: 2200, epoch: 9 | loss: 0.3003105 speed: 0.0710s/iter; left time: 169.3892s Epoch: 9 cost time: 169.20373845100403 Epoch: 9, Steps: 2293 | Train Loss: 0.5863524 Vali Loss: 0.5405004 Test Loss: 0.2119147 Validation loss decreased (0.541890 --> 0.540500). Saving model ... Updating learning rate to 1.953125e-06 iters: 100, epoch: 10 | loss: 0.5354117 speed: 1.0675s/iter; left time: 2342.1284s iters: 200, epoch: 10 | loss: 0.3361762 speed: 0.0721s/iter; left time: 150.8782s iters: 300, epoch: 10 | loss: 0.5681809 speed: 0.0719s/iter; left time: 143.3719s iters: 400, epoch: 10 | loss: 2.3432953 speed: 0.0728s/iter; left time: 137.9589s iters: 500, epoch: 10 | loss: 0.6368713 speed: 0.0691s/iter; left time: 124.0376s iters: 600, epoch: 10 | loss: 0.4658722 speed: 0.0781s/iter; left time: 132.2914s iters: 700, epoch: 10 | loss: 2.3447948 speed: 0.0744s/iter; left time: 118.6382s iters: 800, epoch: 10 | loss: 0.3918543 speed: 0.0723s/iter; left time: 107.9719s iters: 900, epoch: 10 | loss: 0.8322681 speed: 0.0765s/iter; left time: 106.6387s iters: 1000, epoch: 10 | loss: 0.5292996 speed: 0.0762s/iter; left time: 98.5937s iters: 1100, epoch: 10 | loss: 0.4620808 speed: 0.0726s/iter; left time: 86.6297s iters: 1200, epoch: 10 | loss: 0.4627773 speed: 0.0714s/iter; left time: 78.1347s iters: 1300, epoch: 10 | loss: 0.6950729 speed: 0.0845s/iter; left time: 83.9461s iters: 1400, epoch: 10 | loss: 0.5145064 speed: 0.0767s/iter; left time: 68.5525s iters: 1500, epoch: 10 | loss: 0.5445551 speed: 0.0742s/iter; left time: 58.8932s iters: 1600, epoch: 10 | loss: 2.4414353 speed: 0.0734s/iter; left time: 50.9703s iters: 1700, epoch: 10 | loss: 0.7854645 speed: 0.0753s/iter; left time: 44.7192s iters: 1800, epoch: 10 | loss: 0.6586080 speed: 0.0743s/iter; left time: 36.7134s iters: 1900, epoch: 10 | loss: 0.4719872 speed: 0.0767s/iter; left time: 30.2148s iters: 2000, epoch: 10 | loss: 0.4536229 speed: 0.0732s/iter; left time: 21.5276s iters: 2100, epoch: 10 | loss: 0.5223879 speed: 0.0719s/iter; left time: 13.9514s iters: 2200, epoch: 10 | loss: 0.3328162 speed: 0.0734s/iter; left time: 6.9032s Epoch: 10 cost time: 171.80304598808289 Epoch: 10, Steps: 2293 | Train Loss: 0.5863538 Vali Loss: 0.5392951 Test Loss: 0.2118448 Validation loss decreased (0.540500 --> 0.539295). Saving model ... Updating learning rate to 9.765625e-07 testing : long_term_forecast_ECL_96_iTransformer_custom_ftM_sl96_ll48_pl96_dm512_nh8_el3_dl1_df512_fc3_ebtimeF_dtTrue_test_0<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<< test 10444 test shape: (10444, 1, 96, 21) (10444, 1, 96, 21) test shape: (10444, 96, 21) (10444, 96, 21) mse:0.21184469759464264, mae:0.2560219466686249
Process finished with exit code 0
您好,请参考iTransformer的官方代码:https://github.com/thuml/iTransformer
ssh://hugh@202.117.21.53:22/home/hugh/anaconda3/bin/python -u /home/share/hugh/project/timesnet/run_itrans.py Args in experiment: Namespace(activation='gelu', anomaly_ratio=0.25, batch_size=16, c_out=321, checkpoints='./checkpoints/', d_ff=512, d_layers=1, d_model=512, data='custom', data_path='weather.csv', dec_in=321, des='test', devices='0,1,2,3', distil=True, dropout=0.1, e_layers=3, embed='timeF', enc_in=321, factor=3, features='M', freq='h', gpu=0, inverse=False, is_training=1, itr=1, label_len=48, learning_rate=0.0005, loss='MSE', lradj='type1', mask_rate=0.25, model='iTransformer', model_id='ECL_96', moving_avg=25, n_heads=8, num_kernels=6, num_workers=10, output_attention=False, p_hidden_dims=[128, 128], p_hidden_layers=2, patience=3, pred_len=96, root_path='./dataset/', seasonal_patterns='Monthly', seq_len=96, target='OT', task_name='long_term_forecast', top_k=5, train_epochs=10, use_amp=False, use_gpu=True, use_multi_gpu=False) Use GPU: cuda:0 Number of parameters: 4.83M
Process finished with exit code 0