<!-- 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. -->
20230829234145
This model is a fine-tuned version of bert-large-cased on the super_glue dataset. It achieves the following results on the evaluation set:
- Loss: 0.5907
- Accuracy: 0.6635
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: 0.007
- train_batch_size: 16
- eval_batch_size: 8
- seed: 44
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 80.0
Training results
Training Loss | Epoch | Step | Validation Loss | Accuracy |
---|---|---|---|---|
No log | 1.0 | 35 | 0.6230 | 0.5192 |
No log | 2.0 | 70 | 0.7246 | 0.4135 |
No log | 3.0 | 105 | 0.9519 | 0.5962 |
No log | 4.0 | 140 | 0.8762 | 0.6346 |
No log | 5.0 | 175 | 1.0190 | 0.3558 |
No log | 6.0 | 210 | 0.7460 | 0.625 |
No log | 7.0 | 245 | 0.9006 | 0.4038 |
No log | 8.0 | 280 | 0.6289 | 0.5 |
No log | 9.0 | 315 | 1.2662 | 0.3654 |
No log | 10.0 | 350 | 0.7414 | 0.4327 |
No log | 11.0 | 385 | 0.6525 | 0.4231 |
No log | 12.0 | 420 | 0.6524 | 0.5769 |
No log | 13.0 | 455 | 1.9532 | 0.3654 |
No log | 14.0 | 490 | 1.1259 | 0.6346 |
0.9681 | 15.0 | 525 | 0.5842 | 0.6346 |
0.9681 | 16.0 | 560 | 0.6605 | 0.6346 |
0.9681 | 17.0 | 595 | 0.7591 | 0.3942 |
0.9681 | 18.0 | 630 | 0.5935 | 0.5865 |
0.9681 | 19.0 | 665 | 0.5999 | 0.5865 |
0.9681 | 20.0 | 700 | 0.5997 | 0.625 |
0.9681 | 21.0 | 735 | 0.6639 | 0.6346 |
0.9681 | 22.0 | 770 | 0.6340 | 0.4808 |
0.9681 | 23.0 | 805 | 0.7496 | 0.3654 |
0.9681 | 24.0 | 840 | 0.6882 | 0.4135 |
0.9681 | 25.0 | 875 | 0.8965 | 0.375 |
0.9681 | 26.0 | 910 | 0.6820 | 0.4231 |
0.9681 | 27.0 | 945 | 0.6356 | 0.6346 |
0.9681 | 28.0 | 980 | 0.6703 | 0.3558 |
0.809 | 29.0 | 1015 | 0.6725 | 0.3654 |
0.809 | 30.0 | 1050 | 1.0256 | 0.3654 |
0.809 | 31.0 | 1085 | 1.0189 | 0.3654 |
0.809 | 32.0 | 1120 | 0.6157 | 0.5192 |
0.809 | 33.0 | 1155 | 0.7614 | 0.6346 |
0.809 | 34.0 | 1190 | 0.5927 | 0.5962 |
0.809 | 35.0 | 1225 | 1.7376 | 0.3654 |
0.809 | 36.0 | 1260 | 0.6220 | 0.5288 |
0.809 | 37.0 | 1295 | 1.1171 | 0.6346 |
0.809 | 38.0 | 1330 | 0.6991 | 0.6346 |
0.809 | 39.0 | 1365 | 0.7000 | 0.3942 |
0.809 | 40.0 | 1400 | 0.6723 | 0.4135 |
0.809 | 41.0 | 1435 | 0.9776 | 0.3654 |
0.809 | 42.0 | 1470 | 0.7682 | 0.3654 |
0.8083 | 43.0 | 1505 | 0.5973 | 0.6346 |
0.8083 | 44.0 | 1540 | 0.6068 | 0.6346 |
0.8083 | 45.0 | 1575 | 0.7551 | 0.3654 |
0.8083 | 46.0 | 1610 | 0.5952 | 0.6154 |
0.8083 | 47.0 | 1645 | 0.5828 | 0.6346 |
0.8083 | 48.0 | 1680 | 0.5800 | 0.6346 |
0.8083 | 49.0 | 1715 | 0.5863 | 0.6346 |
0.8083 | 50.0 | 1750 | 0.6166 | 0.5 |
0.8083 | 51.0 | 1785 | 0.6967 | 0.4231 |
0.8083 | 52.0 | 1820 | 0.7029 | 0.3942 |
0.8083 | 53.0 | 1855 | 0.9476 | 0.3654 |
0.8083 | 54.0 | 1890 | 0.8069 | 0.3942 |
0.8083 | 55.0 | 1925 | 0.5984 | 0.6346 |
0.8083 | 56.0 | 1960 | 0.5889 | 0.6346 |
0.8083 | 57.0 | 1995 | 0.6608 | 0.3942 |
0.7064 | 58.0 | 2030 | 0.6557 | 0.4038 |
0.7064 | 59.0 | 2065 | 0.5971 | 0.6346 |
0.7064 | 60.0 | 2100 | 0.6095 | 0.6346 |
0.7064 | 61.0 | 2135 | 0.6373 | 0.6346 |
0.7064 | 62.0 | 2170 | 0.6203 | 0.4423 |
0.7064 | 63.0 | 2205 | 0.6025 | 0.5865 |
0.7064 | 64.0 | 2240 | 0.7393 | 0.6346 |
0.7064 | 65.0 | 2275 | 0.5843 | 0.6346 |
0.7064 | 66.0 | 2310 | 0.6253 | 0.4327 |
0.7064 | 67.0 | 2345 | 0.5865 | 0.6346 |
0.7064 | 68.0 | 2380 | 0.6584 | 0.4327 |
0.7064 | 69.0 | 2415 | 0.6112 | 0.6346 |
0.7064 | 70.0 | 2450 | 0.6089 | 0.6346 |
0.7064 | 71.0 | 2485 | 0.5796 | 0.6538 |
0.6752 | 72.0 | 2520 | 0.6078 | 0.5481 |
0.6752 | 73.0 | 2555 | 0.5944 | 0.5865 |
0.6752 | 74.0 | 2590 | 0.6321 | 0.4519 |
0.6752 | 75.0 | 2625 | 0.5994 | 0.5577 |
0.6752 | 76.0 | 2660 | 0.5935 | 0.625 |
0.6752 | 77.0 | 2695 | 0.7270 | 0.3846 |
0.6752 | 78.0 | 2730 | 0.6153 | 0.5288 |
0.6752 | 79.0 | 2765 | 0.5910 | 0.6635 |
0.6752 | 80.0 | 2800 | 0.5907 | 0.6635 |
Framework versions
- Transformers 4.26.1
- Pytorch 2.0.1+cu118
- Datasets 2.12.0
- Tokenizers 0.13.3