getting 38% hamming score using this code for multilabel classification of my 24000 dataset into 26 classes. please suggest something so that casn go upto 90%

[anushka192001]

0it [00:00, ?it/s]
Epoch: 0, Loss: 0.607304036617279
102it [01:51, 1.04s/it]
Epoch: 0, Loss: 0.42913737893104553
202it [03:40, 1.04s/it]
Epoch: 0, Loss: 0.31689316034317017
278it [05:05, 1.10s/it]
0it [00:00, ?it/s]
Epoch: 1, Loss: 0.25217771530151367
102it [01:52, 1.04s/it]
Epoch: 1, Loss: 0.2359163910150528
202it [03:41, 1.04s/it]
Epoch: 1, Loss: 0.20691940188407898
278it [05:06, 1.10s/it]
0it [00:00, ?it/s]
Epoch: 2, Loss: 0.18726885318756104
102it [01:52, 1.05s/it]
Epoch: 2, Loss: 0.18049341440200806
202it [03:41, 1.05s/it]
Epoch: 2, Loss: 0.1675453633069992
278it [05:06, 1.10s/it]
0it [00:00, ?it/s]
Epoch: 3, Loss: 0.1804620772600174
102it [01:52, 1.04s/it]
Epoch: 3, Loss: 0.15942974388599396
202it [03:41, 1.04s/it]
Epoch: 3, Loss: 0.1614547073841095
278it [05:06, 1.10s/it]
0it [00:00, ?it/s]
Epoch: 4, Loss: 0.16714756190776825
102it [01:52, 1.04s/it]
Epoch: 4, Loss: 0.18206024169921875
202it [03:41, 1.04s/it]
Epoch: 4, Loss: 0.18216757476329803
278it [05:06, 1.10s/it]
0it [00:00, ?it/s]
Epoch: 5, Loss: 0.14875689148902893
102it [01:52, 1.03s/it]
Epoch: 5, Loss: 0.14692261815071106
202it [03:40, 1.04s/it]
Epoch: 5, Loss: 0.1404387503862381
278it [05:04, 1.10s/it]
0it [00:00, ?it/s]
Epoch: 6, Loss: 0.16587977111339