IMDB_ELECTRA_5E
This model is a fine-tuned version of google/electra-base-discriminator on the imdb dataset. It achieves the following results on the evaluation set:
- Loss: 0.2158
- Accuracy: 0.9533
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 1e-05
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 5
Training results
Training Loss | Epoch | Step | Validation Loss | Accuracy |
---|---|---|---|---|
0.6784 | 0.03 | 50 | 0.6027 | 0.84 |
0.4378 | 0.06 | 100 | 0.2217 | 0.9533 |
0.3063 | 0.1 | 150 | 0.1879 | 0.94 |
0.2183 | 0.13 | 200 | 0.1868 | 0.9333 |
0.2058 | 0.16 | 250 | 0.1548 | 0.9467 |
0.2287 | 0.19 | 300 | 0.1572 | 0.9533 |
0.1677 | 0.22 | 350 | 0.1472 | 0.9533 |
0.1997 | 0.26 | 400 | 0.1400 | 0.9533 |
0.1966 | 0.29 | 450 | 0.1592 | 0.9333 |
0.184 | 0.32 | 500 | 0.1529 | 0.96 |
0.2084 | 0.35 | 550 | 0.2136 | 0.9067 |
0.2126 | 0.38 | 600 | 0.1508 | 0.9533 |
0.1695 | 0.42 | 650 | 0.1442 | 0.9667 |
0.2506 | 0.45 | 700 | 0.1811 | 0.9467 |
0.1755 | 0.48 | 750 | 0.1336 | 0.96 |
0.1874 | 0.51 | 800 | 0.1403 | 0.9533 |
0.1535 | 0.54 | 850 | 0.1239 | 0.96 |
0.1458 | 0.58 | 900 | 0.1198 | 0.9667 |
0.1649 | 0.61 | 950 | 0.1538 | 0.9533 |
0.2014 | 0.64 | 1000 | 0.1196 | 0.9667 |
0.1651 | 0.67 | 1050 | 0.1200 | 0.96 |
0.1595 | 0.7 | 1100 | 0.1155 | 0.96 |
0.1787 | 0.74 | 1150 | 0.1175 | 0.96 |
0.1666 | 0.77 | 1200 | 0.1264 | 0.9533 |
0.1412 | 0.8 | 1250 | 0.1655 | 0.9533 |
0.1949 | 0.83 | 1300 | 0.1363 | 0.9467 |
0.1485 | 0.86 | 1350 | 0.1434 | 0.9667 |
0.1801 | 0.9 | 1400 | 0.1379 | 0.9667 |
0.178 | 0.93 | 1450 | 0.1498 | 0.96 |
0.1767 | 0.96 | 1500 | 0.1507 | 0.9533 |
0.1452 | 0.99 | 1550 | 0.1340 | 0.94 |
0.1465 | 1.02 | 1600 | 0.1416 | 0.96 |
0.1115 | 1.06 | 1650 | 0.1435 | 0.96 |
0.1212 | 1.09 | 1700 | 0.1379 | 0.96 |
0.1534 | 1.12 | 1750 | 0.1198 | 0.96 |
0.1164 | 1.15 | 1800 | 0.1011 | 0.96 |
0.1383 | 1.18 | 1850 | 0.1043 | 0.96 |
0.1415 | 1.22 | 1900 | 0.0914 | 0.9533 |
0.136 | 1.25 | 1950 | 0.1341 | 0.9533 |
0.1301 | 1.28 | 2000 | 0.1303 | 0.9533 |
0.1486 | 1.31 | 2050 | 0.1027 | 0.9733 |
0.0844 | 1.34 | 2100 | 0.1410 | 0.9667 |
0.1388 | 1.38 | 2150 | 0.1265 | 0.9667 |
0.12 | 1.41 | 2200 | 0.1139 | 0.9667 |
0.1329 | 1.44 | 2250 | 0.1259 | 0.9667 |
0.0982 | 1.47 | 2300 | 0.1349 | 0.9667 |
0.1271 | 1.5 | 2350 | 0.1176 | 0.9667 |
0.1286 | 1.54 | 2400 | 0.1349 | 0.9533 |
0.1079 | 1.57 | 2450 | 0.1335 | 0.96 |
0.1236 | 1.6 | 2500 | 0.1393 | 0.96 |
0.1285 | 1.63 | 2550 | 0.1635 | 0.96 |
0.0932 | 1.66 | 2600 | 0.1571 | 0.9533 |
0.1222 | 1.7 | 2650 | 0.1610 | 0.9533 |
0.1421 | 1.73 | 2700 | 0.1296 | 0.9533 |
0.1581 | 1.76 | 2750 | 0.1289 | 0.96 |
0.1245 | 1.79 | 2800 | 0.1180 | 0.9667 |
0.1196 | 1.82 | 2850 | 0.1371 | 0.96 |
0.1062 | 1.86 | 2900 | 0.1269 | 0.96 |
0.1188 | 1.89 | 2950 | 0.1259 | 0.9667 |
0.1183 | 1.92 | 3000 | 0.1164 | 0.9667 |
0.1173 | 1.95 | 3050 | 0.1280 | 0.9667 |
0.1344 | 1.98 | 3100 | 0.1439 | 0.96 |
0.1166 | 2.02 | 3150 | 0.1442 | 0.96 |
0.0746 | 2.05 | 3200 | 0.1562 | 0.96 |
0.0813 | 2.08 | 3250 | 0.1760 | 0.96 |
0.0991 | 2.11 | 3300 | 0.1485 | 0.9667 |
0.076 | 2.14 | 3350 | 0.1530 | 0.9533 |
0.087 | 2.18 | 3400 | 0.1441 | 0.96 |
0.0754 | 2.21 | 3450 | 0.1401 | 0.9667 |
0.0878 | 2.24 | 3500 | 0.1480 | 0.96 |
0.0605 | 2.27 | 3550 | 0.1579 | 0.9667 |
0.0424 | 2.3 | 3600 | 0.1897 | 0.9667 |
0.0541 | 2.34 | 3650 | 0.1784 | 0.96 |
0.0755 | 2.37 | 3700 | 0.1527 | 0.9733 |
0.1089 | 2.4 | 3750 | 0.1376 | 0.9733 |
0.1061 | 2.43 | 3800 | 0.1329 | 0.9667 |
0.0858 | 2.46 | 3850 | 0.1539 | 0.9667 |
0.1424 | 2.5 | 3900 | 0.1296 | 0.9667 |
0.0928 | 2.53 | 3950 | 0.1324 | 0.9667 |
0.0669 | 2.56 | 4000 | 0.1371 | 0.9667 |
0.0797 | 2.59 | 4050 | 0.1493 | 0.9667 |
0.0563 | 2.62 | 4100 | 0.1657 | 0.96 |
0.0579 | 2.66 | 4150 | 0.1799 | 0.9533 |
0.1014 | 2.69 | 4200 | 0.1625 | 0.96 |
0.0629 | 2.72 | 4250 | 0.1388 | 0.9733 |
0.1331 | 2.75 | 4300 | 0.1522 | 0.9667 |
0.0535 | 2.78 | 4350 | 0.1449 | 0.9667 |
0.1103 | 2.82 | 4400 | 0.1394 | 0.9733 |
0.0691 | 2.85 | 4450 | 0.1324 | 0.9733 |
0.0869 | 2.88 | 4500 | 0.1146 | 0.9667 |
0.068 | 2.91 | 4550 | 0.1621 | 0.9667 |
0.0854 | 2.94 | 4600 | 0.1995 | 0.96 |
0.0907 | 2.98 | 4650 | 0.1819 | 0.96 |
0.0679 | 3.01 | 4700 | 0.1771 | 0.9533 |
0.0632 | 3.04 | 4750 | 0.1388 | 0.9667 |
0.0653 | 3.07 | 4800 | 0.1652 | 0.9667 |
0.0305 | 3.1 | 4850 | 0.1474 | 0.9733 |
0.065 | 3.13 | 4900 | 0.1741 | 0.9667 |
0.0909 | 3.17 | 4950 | 0.1417 | 0.9733 |
0.0663 | 3.2 | 5000 | 0.1578 | 0.9667 |
0.0204 | 3.23 | 5050 | 0.1801 | 0.9667 |
0.0478 | 3.26 | 5100 | 0.1892 | 0.9667 |
0.0809 | 3.29 | 5150 | 0.1724 | 0.9667 |
0.0454 | 3.33 | 5200 | 0.2045 | 0.96 |
0.0958 | 3.36 | 5250 | 0.1635 | 0.9667 |
0.0258 | 3.39 | 5300 | 0.1831 | 0.9667 |
0.0621 | 3.42 | 5350 | 0.1663 | 0.9667 |
0.064 | 3.45 | 5400 | 0.1794 | 0.9667 |
0.0629 | 3.49 | 5450 | 0.1737 | 0.9667 |
0.0436 | 3.52 | 5500 | 0.1815 | 0.9667 |
0.0378 | 3.55 | 5550 | 0.1903 | 0.9667 |
0.0149 | 3.58 | 5600 | 0.1876 | 0.9667 |
0.0698 | 3.61 | 5650 | 0.1861 | 0.9667 |
0.047 | 3.65 | 5700 | 0.1764 | 0.9667 |
0.0739 | 3.68 | 5750 | 0.1510 | 0.9667 |
0.0363 | 3.71 | 5800 | 0.1802 | 0.96 |
0.031 | 3.74 | 5850 | 0.1688 | 0.9733 |
0.1034 | 3.77 | 5900 | 0.1764 | 0.9667 |
0.0588 | 3.81 | 5950 | 0.1840 | 0.9667 |
0.0433 | 3.84 | 6000 | 0.1743 | 0.9667 |
0.057 | 3.87 | 6050 | 0.1896 | 0.9667 |
0.0385 | 3.9 | 6100 | 0.1959 | 0.9667 |
0.0483 | 3.93 | 6150 | 0.1982 | 0.9667 |
0.0292 | 3.97 | 6200 | 0.2016 | 0.9667 |
0.0456 | 4.0 | 6250 | 0.1981 | 0.9667 |
0.0563 | 4.03 | 6300 | 0.1915 | 0.9667 |
0.0346 | 4.06 | 6350 | 0.1967 | 0.9667 |
0.038 | 4.09 | 6400 | 0.2035 | 0.9667 |
0.0341 | 4.13 | 6450 | 0.2356 | 0.96 |
0.0425 | 4.16 | 6500 | 0.1913 | 0.9667 |
0.0282 | 4.19 | 6550 | 0.2091 | 0.96 |
0.0543 | 4.22 | 6600 | 0.2311 | 0.9533 |
0.0139 | 4.25 | 6650 | 0.2260 | 0.96 |
0.0587 | 4.29 | 6700 | 0.2257 | 0.96 |
0.0446 | 4.32 | 6750 | 0.2439 | 0.9533 |
0.0447 | 4.35 | 6800 | 0.2444 | 0.9533 |
0.0199 | 4.38 | 6850 | 0.2327 | 0.96 |
0.0392 | 4.41 | 6900 | 0.2476 | 0.9533 |
0.0596 | 4.45 | 6950 | 0.2443 | 0.9533 |
0.0292 | 4.48 | 7000 | 0.2499 | 0.9533 |
0.0325 | 4.51 | 7050 | 0.2430 | 0.9533 |
0.0316 | 4.54 | 7100 | 0.2272 | 0.96 |
0.0259 | 4.57 | 7150 | 0.2275 | 0.96 |
0.0294 | 4.61 | 7200 | 0.2339 | 0.9533 |
0.0292 | 4.64 | 7250 | 0.2304 | 0.96 |
0.0258 | 4.67 | 7300 | 0.2258 | 0.96 |
0.0221 | 4.7 | 7350 | 0.2164 | 0.96 |
0.0407 | 4.73 | 7400 | 0.2212 | 0.96 |
0.0344 | 4.77 | 7450 | 0.2246 | 0.96 |
0.0364 | 4.8 | 7500 | 0.2211 | 0.96 |
0.0266 | 4.83 | 7550 | 0.2207 | 0.96 |
0.0419 | 4.86 | 7600 | 0.2199 | 0.9533 |
0.0283 | 4.89 | 7650 | 0.2185 | 0.96 |
0.0193 | 4.93 | 7700 | 0.2173 | 0.96 |
0.022 | 4.96 | 7750 | 0.2157 | 0.9533 |
0.0517 | 4.99 | 7800 | 0.2158 | 0.9533 |
Framework versions
- Transformers 4.23.1
- Pytorch 1.13.0
- Datasets 2.6.1
- Tokenizers 0.13.1
- Downloads last month
- 3