t5-small_stereoset_finetuned
This model is a fine-tuned version of t5-small on the stereoset dataset. It achieves the following results on the evaluation set:
- Loss: 0.3572
- Accuracy: 0.5181
- Tp: 0.5008
- Tn: 0.0173
- Fp: 0.4819
- Fn: 0.0
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: 5e-05
- train_batch_size: 64
- eval_batch_size: 64
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 50
Training results
Training Loss | Epoch | Step | Validation Loss | Accuracy | Tp | Tn | Fp | Fn |
---|---|---|---|---|---|---|---|---|
6.2515 | 0.43 | 20 | 5.1550 | 0.6507 | 0.4992 | 0.1515 | 0.3477 | 0.0016 |
3.3425 | 0.85 | 40 | 1.9895 | 0.8414 | 0.4451 | 0.3964 | 0.1028 | 0.0557 |
1.0076 | 1.28 | 60 | 0.5623 | 0.5110 | 0.5 | 0.0110 | 0.4882 | 0.0008 |
0.522 | 1.7 | 80 | 0.3896 | 0.5008 | 0.5008 | 0.0 | 0.4992 | 0.0 |
0.4541 | 2.13 | 100 | 0.3777 | 0.5008 | 0.5008 | 0.0 | 0.4992 | 0.0 |
0.4494 | 2.55 | 120 | 0.3694 | 0.5008 | 0.5008 | 0.0 | 0.4992 | 0.0 |
0.3915 | 2.98 | 140 | 0.3468 | 0.5008 | 0.5008 | 0.0 | 0.4992 | 0.0 |
0.3956 | 3.4 | 160 | 0.3510 | 0.5008 | 0.5008 | 0.0 | 0.4992 | 0.0 |
0.395 | 3.83 | 180 | 0.3403 | 0.5008 | 0.5008 | 0.0 | 0.4992 | 0.0 |
0.4005 | 4.26 | 200 | 0.3357 | 0.5008 | 0.5008 | 0.0 | 0.4992 | 0.0 |
0.3764 | 4.68 | 220 | 0.3367 | 0.5008 | 0.5008 | 0.0 | 0.4992 | 0.0 |
0.3798 | 5.11 | 240 | 0.3321 | 0.5008 | 0.5008 | 0.0 | 0.4992 | 0.0 |
0.3546 | 5.53 | 260 | 0.3452 | 0.5008 | 0.5008 | 0.0 | 0.4992 | 0.0 |
0.3716 | 5.96 | 280 | 0.3257 | 0.5008 | 0.5008 | 0.0 | 0.4992 | 0.0 |
0.3598 | 6.38 | 300 | 0.3255 | 0.5008 | 0.5008 | 0.0 | 0.4992 | 0.0 |
0.366 | 6.81 | 320 | 0.3400 | 0.5008 | 0.5008 | 0.0 | 0.4992 | 0.0 |
0.3537 | 7.23 | 340 | 0.3239 | 0.5008 | 0.5008 | 0.0 | 0.4992 | 0.0 |
0.3395 | 7.66 | 360 | 0.3238 | 0.5008 | 0.5008 | 0.0 | 0.4992 | 0.0 |
0.338 | 8.09 | 380 | 0.3323 | 0.5008 | 0.5008 | 0.0 | 0.4992 | 0.0 |
0.3402 | 8.51 | 400 | 0.3193 | 0.5008 | 0.5008 | 0.0 | 0.4992 | 0.0 |
0.3421 | 8.94 | 420 | 0.3188 | 0.5008 | 0.5008 | 0.0 | 0.4992 | 0.0 |
0.318 | 9.36 | 440 | 0.3173 | 0.5008 | 0.5008 | 0.0 | 0.4992 | 0.0 |
0.3532 | 9.79 | 460 | 0.3164 | 0.5008 | 0.5008 | 0.0 | 0.4992 | 0.0 |
0.3333 | 10.21 | 480 | 0.3175 | 0.5008 | 0.5008 | 0.0 | 0.4992 | 0.0 |
0.3187 | 10.64 | 500 | 0.3157 | 0.5008 | 0.5008 | 0.0 | 0.4992 | 0.0 |
0.3004 | 11.06 | 520 | 0.3108 | 0.5008 | 0.5008 | 0.0 | 0.4992 | 0.0 |
0.3257 | 11.49 | 540 | 0.3094 | 0.5008 | 0.5008 | 0.0 | 0.4992 | 0.0 |
0.3342 | 11.91 | 560 | 0.3125 | 0.5008 | 0.5008 | 0.0 | 0.4992 | 0.0 |
0.3084 | 12.34 | 580 | 0.3141 | 0.5008 | 0.5008 | 0.0 | 0.4992 | 0.0 |
0.3134 | 12.77 | 600 | 0.3076 | 0.5008 | 0.5008 | 0.0 | 0.4992 | 0.0 |
0.3081 | 13.19 | 620 | 0.3059 | 0.5008 | 0.5008 | 0.0 | 0.4992 | 0.0 |
0.3017 | 13.62 | 640 | 0.3098 | 0.5008 | 0.5008 | 0.0 | 0.4992 | 0.0 |
0.2924 | 14.04 | 660 | 0.3046 | 0.5008 | 0.5008 | 0.0 | 0.4992 | 0.0 |
0.3109 | 14.47 | 680 | 0.3054 | 0.5008 | 0.5008 | 0.0 | 0.4992 | 0.0 |
0.287 | 14.89 | 700 | 0.3061 | 0.5008 | 0.5008 | 0.0 | 0.4992 | 0.0 |
0.2783 | 15.32 | 720 | 0.3051 | 0.5008 | 0.5008 | 0.0 | 0.4992 | 0.0 |
0.2938 | 15.74 | 740 | 0.3037 | 0.5031 | 0.5008 | 0.0024 | 0.4969 | 0.0 |
0.2788 | 16.17 | 760 | 0.3057 | 0.5024 | 0.5008 | 0.0016 | 0.4976 | 0.0 |
0.2872 | 16.6 | 780 | 0.3100 | 0.5016 | 0.5008 | 0.0008 | 0.4984 | 0.0 |
0.2794 | 17.02 | 800 | 0.3055 | 0.5031 | 0.5008 | 0.0024 | 0.4969 | 0.0 |
0.2847 | 17.45 | 820 | 0.3047 | 0.5055 | 0.5008 | 0.0047 | 0.4945 | 0.0 |
0.2644 | 17.87 | 840 | 0.3067 | 0.5024 | 0.5008 | 0.0016 | 0.4976 | 0.0 |
0.2558 | 18.3 | 860 | 0.3062 | 0.5063 | 0.5008 | 0.0055 | 0.4937 | 0.0 |
0.2867 | 18.72 | 880 | 0.3160 | 0.5024 | 0.5008 | 0.0016 | 0.4976 | 0.0 |
0.2864 | 19.15 | 900 | 0.3067 | 0.5071 | 0.5008 | 0.0063 | 0.4929 | 0.0 |
0.2645 | 19.57 | 920 | 0.3211 | 0.5024 | 0.5008 | 0.0016 | 0.4976 | 0.0 |
0.2606 | 20.0 | 940 | 0.3067 | 0.5094 | 0.5008 | 0.0086 | 0.4906 | 0.0 |
0.2694 | 20.43 | 960 | 0.3125 | 0.5055 | 0.5008 | 0.0047 | 0.4945 | 0.0 |
0.2634 | 20.85 | 980 | 0.3072 | 0.5086 | 0.5008 | 0.0078 | 0.4914 | 0.0 |
0.2519 | 21.28 | 1000 | 0.3088 | 0.5086 | 0.5008 | 0.0078 | 0.4914 | 0.0 |
0.2537 | 21.7 | 1020 | 0.3136 | 0.5071 | 0.5008 | 0.0063 | 0.4929 | 0.0 |
0.2536 | 22.13 | 1040 | 0.3089 | 0.5133 | 0.5008 | 0.0126 | 0.4867 | 0.0 |
0.2488 | 22.55 | 1060 | 0.3108 | 0.5110 | 0.5008 | 0.0102 | 0.4890 | 0.0 |
0.2558 | 22.98 | 1080 | 0.3157 | 0.5071 | 0.5008 | 0.0063 | 0.4929 | 0.0 |
0.2626 | 23.4 | 1100 | 0.3104 | 0.5133 | 0.5008 | 0.0126 | 0.4867 | 0.0 |
0.2424 | 23.83 | 1120 | 0.3130 | 0.5133 | 0.5008 | 0.0126 | 0.4867 | 0.0 |
0.2496 | 24.26 | 1140 | 0.3139 | 0.5133 | 0.5008 | 0.0126 | 0.4867 | 0.0 |
0.2451 | 24.68 | 1160 | 0.3160 | 0.5133 | 0.5008 | 0.0126 | 0.4867 | 0.0 |
0.2399 | 25.11 | 1180 | 0.3148 | 0.5141 | 0.5008 | 0.0133 | 0.4859 | 0.0 |
0.236 | 25.53 | 1200 | 0.3163 | 0.5141 | 0.5008 | 0.0133 | 0.4859 | 0.0 |
0.2561 | 25.96 | 1220 | 0.3208 | 0.5118 | 0.5008 | 0.0110 | 0.4882 | 0.0 |
0.2425 | 26.38 | 1240 | 0.3161 | 0.5149 | 0.5008 | 0.0141 | 0.4851 | 0.0 |
0.2376 | 26.81 | 1260 | 0.3173 | 0.5133 | 0.5008 | 0.0126 | 0.4867 | 0.0 |
0.246 | 27.23 | 1280 | 0.3174 | 0.5149 | 0.5008 | 0.0141 | 0.4851 | 0.0 |
0.2256 | 27.66 | 1300 | 0.3186 | 0.5149 | 0.5008 | 0.0141 | 0.4851 | 0.0 |
0.2316 | 28.09 | 1320 | 0.3177 | 0.5149 | 0.5008 | 0.0141 | 0.4851 | 0.0 |
0.2406 | 28.51 | 1340 | 0.3197 | 0.5157 | 0.5008 | 0.0149 | 0.4843 | 0.0 |
0.218 | 28.94 | 1360 | 0.3227 | 0.5133 | 0.5008 | 0.0126 | 0.4867 | 0.0 |
0.2103 | 29.36 | 1380 | 0.3228 | 0.5141 | 0.5008 | 0.0133 | 0.4859 | 0.0 |
0.2161 | 29.79 | 1400 | 0.3244 | 0.5133 | 0.5008 | 0.0126 | 0.4867 | 0.0 |
0.2172 | 30.21 | 1420 | 0.3246 | 0.5141 | 0.5008 | 0.0133 | 0.4859 | 0.0 |
0.2119 | 30.64 | 1440 | 0.3256 | 0.5141 | 0.5008 | 0.0133 | 0.4859 | 0.0 |
0.2202 | 31.06 | 1460 | 0.3261 | 0.5181 | 0.5008 | 0.0173 | 0.4819 | 0.0 |
0.2244 | 31.49 | 1480 | 0.3297 | 0.5141 | 0.5008 | 0.0133 | 0.4859 | 0.0 |
0.2123 | 31.91 | 1500 | 0.3285 | 0.5165 | 0.5008 | 0.0157 | 0.4835 | 0.0 |
0.1914 | 32.34 | 1520 | 0.3308 | 0.5149 | 0.5008 | 0.0141 | 0.4851 | 0.0 |
0.2338 | 32.77 | 1540 | 0.3311 | 0.5173 | 0.5008 | 0.0165 | 0.4827 | 0.0 |
0.2206 | 33.19 | 1560 | 0.3317 | 0.5173 | 0.5008 | 0.0165 | 0.4827 | 0.0 |
0.2278 | 33.62 | 1580 | 0.3340 | 0.5149 | 0.5008 | 0.0141 | 0.4851 | 0.0 |
0.1982 | 34.04 | 1600 | 0.3343 | 0.5181 | 0.5008 | 0.0173 | 0.4819 | 0.0 |
0.2154 | 34.47 | 1620 | 0.3354 | 0.5141 | 0.5008 | 0.0133 | 0.4859 | 0.0 |
0.1901 | 34.89 | 1640 | 0.3372 | 0.5149 | 0.5008 | 0.0141 | 0.4851 | 0.0 |
0.2087 | 35.32 | 1660 | 0.3396 | 0.5149 | 0.5008 | 0.0141 | 0.4851 | 0.0 |
0.1931 | 35.74 | 1680 | 0.3401 | 0.5173 | 0.5008 | 0.0165 | 0.4827 | 0.0 |
0.1995 | 36.17 | 1700 | 0.3411 | 0.5235 | 0.5008 | 0.0228 | 0.4765 | 0.0 |
0.2226 | 36.6 | 1720 | 0.3413 | 0.5181 | 0.5008 | 0.0173 | 0.4819 | 0.0 |
0.216 | 37.02 | 1740 | 0.3443 | 0.5141 | 0.5008 | 0.0133 | 0.4859 | 0.0 |
0.1911 | 37.45 | 1760 | 0.3434 | 0.5141 | 0.5008 | 0.0133 | 0.4859 | 0.0 |
0.2138 | 37.87 | 1780 | 0.3435 | 0.5173 | 0.5008 | 0.0165 | 0.4827 | 0.0 |
0.1918 | 38.3 | 1800 | 0.3449 | 0.5149 | 0.5008 | 0.0141 | 0.4851 | 0.0 |
0.2021 | 38.72 | 1820 | 0.3462 | 0.5149 | 0.5008 | 0.0141 | 0.4851 | 0.0 |
0.2046 | 39.15 | 1840 | 0.3454 | 0.5149 | 0.5008 | 0.0141 | 0.4851 | 0.0 |
0.2084 | 39.57 | 1860 | 0.3460 | 0.5149 | 0.5008 | 0.0141 | 0.4851 | 0.0 |
0.1834 | 40.0 | 1880 | 0.3462 | 0.5149 | 0.5008 | 0.0141 | 0.4851 | 0.0 |
0.19 | 40.43 | 1900 | 0.3479 | 0.5149 | 0.5008 | 0.0141 | 0.4851 | 0.0 |
0.1877 | 40.85 | 1920 | 0.3479 | 0.5149 | 0.5008 | 0.0141 | 0.4851 | 0.0 |
0.2152 | 41.28 | 1940 | 0.3480 | 0.5212 | 0.5008 | 0.0204 | 0.4788 | 0.0 |
0.1795 | 41.7 | 1960 | 0.3487 | 0.5157 | 0.5008 | 0.0149 | 0.4843 | 0.0 |
0.1804 | 42.13 | 1980 | 0.3502 | 0.5149 | 0.5008 | 0.0141 | 0.4851 | 0.0 |
0.1748 | 42.55 | 2000 | 0.3520 | 0.5149 | 0.5008 | 0.0141 | 0.4851 | 0.0 |
0.177 | 42.98 | 2020 | 0.3518 | 0.5181 | 0.5008 | 0.0173 | 0.4819 | 0.0 |
0.179 | 43.4 | 2040 | 0.3529 | 0.5204 | 0.5008 | 0.0196 | 0.4796 | 0.0 |
0.1895 | 43.83 | 2060 | 0.3538 | 0.5204 | 0.5008 | 0.0196 | 0.4796 | 0.0 |
0.1867 | 44.26 | 2080 | 0.3539 | 0.5196 | 0.5008 | 0.0188 | 0.4804 | 0.0 |
0.2155 | 44.68 | 2100 | 0.3546 | 0.5165 | 0.5008 | 0.0157 | 0.4835 | 0.0 |
0.1783 | 45.11 | 2120 | 0.3550 | 0.5181 | 0.5008 | 0.0173 | 0.4819 | 0.0 |
0.1969 | 45.53 | 2140 | 0.3560 | 0.5157 | 0.5008 | 0.0149 | 0.4843 | 0.0 |
0.1826 | 45.96 | 2160 | 0.3564 | 0.5149 | 0.5008 | 0.0141 | 0.4851 | 0.0 |
0.1957 | 46.38 | 2180 | 0.3571 | 0.5149 | 0.5008 | 0.0141 | 0.4851 | 0.0 |
0.1864 | 46.81 | 2200 | 0.3568 | 0.5157 | 0.5008 | 0.0149 | 0.4843 | 0.0 |
0.1889 | 47.23 | 2220 | 0.3564 | 0.5196 | 0.5008 | 0.0188 | 0.4804 | 0.0 |
0.1837 | 47.66 | 2240 | 0.3569 | 0.5165 | 0.5008 | 0.0157 | 0.4835 | 0.0 |
0.1713 | 48.09 | 2260 | 0.3568 | 0.5196 | 0.5008 | 0.0188 | 0.4804 | 0.0 |
0.1997 | 48.51 | 2280 | 0.3570 | 0.5188 | 0.5008 | 0.0181 | 0.4812 | 0.0 |
0.1844 | 48.94 | 2300 | 0.3570 | 0.5188 | 0.5008 | 0.0181 | 0.4812 | 0.0 |
0.1854 | 49.36 | 2320 | 0.3571 | 0.5181 | 0.5008 | 0.0173 | 0.4819 | 0.0 |
0.1897 | 49.79 | 2340 | 0.3572 | 0.5181 | 0.5008 | 0.0173 | 0.4819 | 0.0 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.13.1
- Datasets 2.10.1
- Tokenizers 0.13.2
- Downloads last month
- 25
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social
visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.
Dataset used to train henryscheible/t5-small_stereoset_finetuned
Evaluation results
- Accuracy on stereosetvalidation set self-reported0.518