config

[pluto-8]

你好，这是单类训练的吗，一次次修改类？

[arimousa]

Hi yes we train a single model for each category. I have provided the checkpoints to avoid extra training. You can find them in readme file.

[pluto-8]

非常感谢

感觉跑的没问题啊怎么总是差一点

[Alide930]

Class: hazelnut w: 2 v: 1 load_chp: 2000 feature extractor: wide_resnet101_2 w_DA: 3 DLlambda: 0.1
config.model.test_trajectoy_steps=250 , config.data.test_batch_size=16
Training...
Num params: 32952707
Epoch 25 | Loss: 723.661865234375
Epoch 50 | Loss: 503.58660888671875
Epoch 75 | Loss: 440.1573486328125
Epoch 100 | Loss: 256.9188537597656
Epoch 125 | Loss: 1187.53662109375
Epoch 150 | Loss: 1018.278076171875
Epoch 175 | Loss: 199.90769958496094
Epoch 200 | Loss: 1030.4249267578125
Epoch 225 | Loss: 4998.15771484375
Epoch 250 | Loss: 436.2159423828125
Epoch 275 | Loss: 447.48187255859375
Epoch 300 | Loss: 126.65884399414062
Epoch 325 | Loss: 341.7905578613281
Epoch 350 | Loss: 240.92970275878906
Epoch 375 | Loss: 929.5860595703125
Epoch 400 | Loss: 2907.0986328125
Epoch 425 | Loss: 241.2455291748047
Epoch 450 | Loss: 149.514404296875
Epoch 475 | Loss: 235.17376708984375
Epoch 500 | Loss: 429.6699523925781
Epoch 525 | Loss: 990.0682373046875
Epoch 550 | Loss: 150.08816528320312
Epoch 575 | Loss: 282.5936279296875
Epoch 600 | Loss: 315.49676513671875
Epoch 625 | Loss: 357.2218017578125
Epoch 650 | Loss: 310.2801208496094
Epoch 675 | Loss: 435.7424621582031
Epoch 700 | Loss: 605.9237060546875
Epoch 725 | Loss: 476.61236572265625
Epoch 750 | Loss: 2133.678955078125
Epoch 775 | Loss: 862.6851196289062
Epoch 800 | Loss: 156.36219787597656
Epoch 825 | Loss: 85.4675064086914
Epoch 850 | Loss: 261.5838623046875
Epoch 875 | Loss: 477.8103942871094
Epoch 900 | Loss: 184.68641662597656
Epoch 925 | Loss: 584.4465942382812
Epoch 950 | Loss: 151.77919006347656
Epoch 975 | Loss: 233.67892456054688
Epoch 1000 | Loss: 1521.38232421875
Epoch 1025 | Loss: 819.031005859375
Epoch 1050 | Loss: 644.1213989257812
Epoch 1075 | Loss: 297.47711181640625
Epoch 1100 | Loss: 890.1820068359375
Epoch 1125 | Loss: 2145.62255859375
Epoch 1150 | Loss: 152.47845458984375
Epoch 1175 | Loss: 237.95982360839844
Epoch 1200 | Loss: 97.16827392578125
Epoch 1225 | Loss: 128.47390747070312
Epoch 1250 | Loss: 319.6527404785156
Epoch 1275 | Loss: 55.759761810302734
Epoch 1300 | Loss: 159.89886474609375
Epoch 1325 | Loss: 942.424560546875
Epoch 1350 | Loss: 91.83270263671875
Epoch 1375 | Loss: 998.3877563476562
Epoch 1400 | Loss: 130.8408203125
Epoch 1425 | Loss: 198.08221435546875
Epoch 1450 | Loss: 155.6403350830078
Epoch 1475 | Loss: 52.975685119628906
Epoch 1500 | Loss: 226.66851806640625
Epoch 1525 | Loss: 271.3487854003906
Epoch 1550 | Loss: 610.568115234375

Hello, was the loss also this high when you were training your model?