CoolFace
Modelpublic

henryscheible/gpt2_stereoset_classifieronly

sourceHugging Facemitupdated 4y agoView on Hugging Face
0likes8downloads
Model Card

<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->

gpt2stereosetclassifieronly

This model is a fine-tuned version of gpt2 on the stereoset dataset. It achieves the following results on the evaluation set:

  • Loss: 0.5990
  • Accuracy: 0.6923
  • Tp: 0.3501
  • Tn: 0.3422
  • Fp: 0.1625
  • Fn: 0.1452

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: 0.0005
  • trainbatchsize: 64
  • evalbatchsize: 64
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lrschedulertype: linear
  • num_epochs: 50

Training results

Training LossEpochStepValidation LossAccuracyTpTnFpFn
0.89220.43200.69130.55490.24020.31480.19000.2551
0.78840.85400.66710.59340.21820.37520.12950.2771
0.69911.28600.65610.61930.22060.39870.10600.2747
0.68191.7800.64990.63110.20880.42230.08240.2865
0.65012.131000.63790.65070.29910.35160.15310.1962
0.65662.551200.65690.61850.16950.44900.05570.3257
0.66712.981400.63130.66090.29430.36660.13810.2009
0.65513.41600.63090.64840.38620.26220.24250.1091
0.6333.831800.62440.66560.30140.36420.14050.1939
0.64324.262000.63200.65540.24020.41520.08950.2551
0.63264.682200.62400.66010.28490.37520.12950.2104
0.63475.112400.62590.65230.36890.28340.22140.1264
0.62045.532600.62560.64990.36970.28020.22450.1256
0.62425.962800.61720.67740.32100.35640.14840.1743
0.61896.383000.61860.65460.34930.30530.19940.1460
0.6256.813200.61870.67270.28810.38460.12010.2072
0.59637.233400.61730.67580.35710.31870.18600.1381
0.62147.663600.61580.66950.32030.34930.15540.1750
0.60078.093800.61230.67970.36110.31870.18600.1342
0.64548.514000.61680.65700.37360.28340.22140.1217
0.60128.944200.61150.68680.33200.35480.14990.1633
0.6279.364400.64850.61930.16560.45370.05100.3297
0.62139.794600.60920.68290.30220.38070.12400.1931
0.628610.214800.61090.67110.36030.31080.19390.1350
0.60910.645000.61340.66330.36110.30220.20250.1342
0.595811.065200.64090.62480.42620.19860.30610.0691
0.649411.495400.63320.63420.41920.21510.28960.0761
0.601211.915600.61590.65930.38850.27080.23390.1068
0.60612.345800.60500.69470.33590.35870.14600.1593
0.587212.776000.61350.66410.38780.27630.22840.1075
0.602613.196200.60610.69620.32650.36970.13500.1688
0.617913.626400.61180.68760.28260.40500.09970.2127
0.574414.046600.60580.69230.30300.38930.11540.1923
0.606114.476800.60720.68600.28490.40110.10360.2104
0.60914.897000.60250.70640.33670.36970.13500.1586
0.601915.327200.60460.68760.35400.33360.17110.1413
0.618315.747400.60870.67350.37910.29430.21040.1162
0.617316.177600.60100.69540.34070.35480.14990.1546
0.587316.67800.60780.67660.38150.29510.20960.1138
0.609517.028000.61510.66250.39480.26770.23700.1005
0.593617.458200.60260.69150.34690.34460.16010.1484
0.582117.878400.60250.69310.34850.34460.16010.1468
0.603618.38600.60320.70490.33910.36580.13890.1562
0.587218.728800.60570.68130.35870.32260.18210.1366
0.608519.159000.60450.68450.35710.32730.17740.1381
0.597219.579200.62030.65620.40420.25200.25270.0911
0.573220.09400.60950.66720.38070.28650.21820.1146
0.571820.439600.60540.68680.29360.39320.11150.2017
0.591920.859800.60310.69310.35010.34300.16170.1452
0.617521.2810000.60880.67030.38230.28810.21660.1130
0.579321.710200.59860.69940.34300.35640.14840.1523
0.594322.1310400.60640.68520.28260.40270.10200.2127
0.571622.5510600.59960.69470.34850.34620.15860.1468
0.611522.9810800.61110.67270.38930.28340.22140.1060
0.598423.411000.60580.68370.38070.30300.20170.1146
0.588223.8311200.59930.69620.33520.36110.14360.1601
0.592424.2611400.61280.66800.39090.27710.22760.1044
0.598424.6811600.60170.69700.32420.37280.13190.1711
0.578125.1111800.60180.70020.33520.36500.13970.1601
0.593725.5312000.60510.68450.36190.32260.18210.1334
0.567825.9612200.59980.70020.32970.37050.13420.1656
0.577626.3812400.62020.65230.39720.25510.24960.0981
0.589126.8112600.60800.68210.37910.30300.20170.1162
0.591527.2312800.60260.69470.29980.39480.10990.1954
0.597227.6613000.59940.69310.35560.33750.16720.1397
0.572128.0913200.60380.68290.37360.30930.19540.1217
0.581328.5113400.59810.69540.33670.35870.14600.1586
0.591428.9413600.59820.69860.33670.36190.14290.1586
0.584829.3613800.59770.70020.33990.36030.14440.1554
0.577229.7914000.60240.68760.36730.32030.18450.1279
0.58130.2114200.60040.69390.36110.33280.17190.1342
0.588130.6414400.59690.70020.34620.35400.15070.1491
0.60131.0614600.59700.69940.33280.36660.13810.1625
0.575931.4914800.59710.69860.33750.36110.14360.1578
0.573831.9115000.59690.70020.34540.35480.14990.1499
0.557632.3415200.59830.69310.34930.34380.16090.1460
0.5832.7715400.59760.70090.33590.36500.13970.1593
0.579833.1915600.59800.70170.34690.35480.14990.1484
0.580233.6215800.59880.69540.34770.34770.15700.1476
0.58734.0416000.59970.69310.35320.33990.16480.1421
0.549934.4716200.60810.67970.38300.29670.20800.1122
0.587834.8916400.59890.69700.34380.35320.15150.1515
0.585535.3216600.60730.68290.38150.30140.20330.1138
0.583635.7416800.59770.70020.33590.36420.14050.1593
0.557636.1717000.59840.69860.33990.35870.14600.1554
0.592936.617200.60350.69070.36970.32100.18370.1256
0.567237.0217400.60230.69230.37050.32180.18290.1248
0.577437.4517600.59860.69470.35090.34380.16090.1444
0.578537.8717800.59900.69620.31950.37680.12790.1758
0.588538.318000.59790.69940.33750.36190.14290.1578
0.544938.7218200.60300.69230.37130.32100.18370.1240
0.585739.1518400.59900.70090.33280.36810.13660.1625
0.583939.5718600.60030.69070.35480.33590.16880.1405
0.580640.018800.59760.69620.34140.35480.14990.1538
0.569240.4319000.59760.70250.33990.36260.14210.1554
0.59340.8519200.59840.69470.34300.35160.15310.1523
0.573641.2819400.59920.69310.35560.33750.16720.1397
0.565341.719600.59780.69700.34380.35320.15150.1515
0.563142.1319800.60060.69470.36030.33440.17030.1350
0.579442.5520000.59830.69940.33360.36580.13890.1617
0.587642.9820200.59840.69390.34220.35160.15310.1531
0.572643.420400.60050.69620.36340.33280.17190.1319
0.56643.8320600.59820.69700.32420.37280.13190.1711
0.560344.2620800.59940.69470.35790.33670.16800.1374
0.569744.6821000.60370.68920.37280.31630.18840.1224
0.562445.1121200.59810.70020.32970.37050.13420.1656
0.564845.5321400.59790.69620.34220.35400.15070.1531
0.57845.9621600.60240.69070.37130.31950.18520.1240
0.559346.3821800.59770.70020.33910.36110.14360.1562
0.575546.8122000.59790.69780.33360.36420.14050.1617
0.5947.2322200.60460.68680.37360.31320.19150.1217
0.564847.6622400.59970.69310.35640.33670.16800.1389
0.581248.0922600.59790.69540.33360.36190.14290.1617
0.579648.5122800.59790.69620.33360.36260.14210.1617
0.570148.9423000.59810.69470.34540.34930.15540.1499
0.580749.3623200.59880.69310.35010.34300.16170.1452
0.583649.7923400.59900.69230.35010.34220.16250.1452

Framework versions

  • Transformers 4.26.1
  • Pytorch 1.13.1
  • Datasets 2.10.1
  • Tokenizers 0.13.2