<!-- 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. -->
20230830020815
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: 0.6170
- Accuracy: 0.5577
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: 0.009
- 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 | 0.8914 | 0.5865 |
No log | 2.0 | 70 | 2.1528 | 0.4135 |
No log | 3.0 | 105 | 1.5770 | 0.4038 |
No log | 4.0 | 140 | 0.8308 | 0.625 |
No log | 5.0 | 175 | 0.7444 | 0.4135 |
No log | 6.0 | 210 | 1.6216 | 0.3654 |
No log | 7.0 | 245 | 0.6072 | 0.6058 |
No log | 8.0 | 280 | 0.9229 | 0.3654 |
No log | 9.0 | 315 | 1.2563 | 0.6346 |
No log | 10.0 | 350 | 1.9376 | 0.3654 |
No log | 11.0 | 385 | 0.8642 | 0.4231 |
No log | 12.0 | 420 | 0.8720 | 0.4135 |
No log | 13.0 | 455 | 0.7846 | 0.625 |
No log | 14.0 | 490 | 1.3697 | 0.6346 |
1.2144 | 15.0 | 525 | 1.2052 | 0.3654 |
1.2144 | 16.0 | 560 | 0.7440 | 0.5962 |
1.2144 | 17.0 | 595 | 0.8147 | 0.5288 |
1.2144 | 18.0 | 630 | 0.8679 | 0.6346 |
1.2144 | 19.0 | 665 | 0.8125 | 0.6346 |
1.2144 | 20.0 | 700 | 0.7545 | 0.5962 |
1.2144 | 21.0 | 735 | 0.5959 | 0.6346 |
1.2144 | 22.0 | 770 | 0.7990 | 0.3654 |
1.2144 | 23.0 | 805 | 0.8083 | 0.6346 |
1.2144 | 24.0 | 840 | 0.6932 | 0.6346 |
1.2144 | 25.0 | 875 | 0.6382 | 0.6346 |
1.2144 | 26.0 | 910 | 0.6023 | 0.625 |
1.2144 | 27.0 | 945 | 0.6502 | 0.625 |
1.2144 | 28.0 | 980 | 0.6646 | 0.4231 |
0.8752 | 29.0 | 1015 | 0.6646 | 0.5288 |
0.8752 | 30.0 | 1050 | 0.6106 | 0.5769 |
0.8752 | 31.0 | 1085 | 0.8355 | 0.375 |
0.8752 | 32.0 | 1120 | 0.6060 | 0.6058 |
0.8752 | 33.0 | 1155 | 0.7944 | 0.375 |
0.8752 | 34.0 | 1190 | 0.6461 | 0.6058 |
0.8752 | 35.0 | 1225 | 0.6320 | 0.5096 |
0.8752 | 36.0 | 1260 | 0.6189 | 0.6154 |
0.8752 | 37.0 | 1295 | 0.6007 | 0.625 |
0.8752 | 38.0 | 1330 | 0.6415 | 0.5096 |
0.8752 | 39.0 | 1365 | 0.6386 | 0.6346 |
0.8752 | 40.0 | 1400 | 0.6051 | 0.5962 |
0.8752 | 41.0 | 1435 | 0.7365 | 0.3942 |
0.8752 | 42.0 | 1470 | 0.7951 | 0.3942 |
0.7496 | 43.0 | 1505 | 0.6346 | 0.5385 |
0.7496 | 44.0 | 1540 | 0.6475 | 0.4712 |
0.7496 | 45.0 | 1575 | 0.7517 | 0.375 |
0.7496 | 46.0 | 1610 | 0.6727 | 0.4327 |
0.7496 | 47.0 | 1645 | 0.6718 | 0.4712 |
0.7496 | 48.0 | 1680 | 0.6113 | 0.5577 |
0.7496 | 49.0 | 1715 | 0.6150 | 0.6346 |
0.7496 | 50.0 | 1750 | 0.6207 | 0.6346 |
0.7496 | 51.0 | 1785 | 0.7305 | 0.375 |
0.7496 | 52.0 | 1820 | 0.5944 | 0.6346 |
0.7496 | 53.0 | 1855 | 0.6348 | 0.4808 |
0.7496 | 54.0 | 1890 | 0.6641 | 0.4808 |
0.7496 | 55.0 | 1925 | 0.6014 | 0.6154 |
0.7496 | 56.0 | 1960 | 0.6118 | 0.6442 |
0.7496 | 57.0 | 1995 | 0.5951 | 0.625 |
0.6833 | 58.0 | 2030 | 0.6069 | 0.5769 |
0.6833 | 59.0 | 2065 | 0.6264 | 0.5865 |
0.6833 | 60.0 | 2100 | 0.6055 | 0.6346 |
0.6833 | 61.0 | 2135 | 0.6010 | 0.6346 |
0.6833 | 62.0 | 2170 | 0.5987 | 0.6154 |
0.6833 | 63.0 | 2205 | 0.6271 | 0.5192 |
0.6833 | 64.0 | 2240 | 0.6102 | 0.6346 |
0.6833 | 65.0 | 2275 | 0.6039 | 0.6058 |
0.6833 | 66.0 | 2310 | 0.6465 | 0.4808 |
0.6833 | 67.0 | 2345 | 0.6219 | 0.5481 |
0.6833 | 68.0 | 2380 | 0.6189 | 0.5481 |
0.6833 | 69.0 | 2415 | 0.5961 | 0.5865 |
0.6833 | 70.0 | 2450 | 0.5996 | 0.6058 |
0.6833 | 71.0 | 2485 | 0.6017 | 0.6058 |
0.6514 | 72.0 | 2520 | 0.6183 | 0.5577 |
0.6514 | 73.0 | 2555 | 0.6026 | 0.5962 |
0.6514 | 74.0 | 2590 | 0.6205 | 0.4808 |
0.6514 | 75.0 | 2625 | 0.6070 | 0.5769 |
0.6514 | 76.0 | 2660 | 0.6173 | 0.4904 |
0.6514 | 77.0 | 2695 | 0.6138 | 0.5385 |
0.6514 | 78.0 | 2730 | 0.6165 | 0.5192 |
0.6514 | 79.0 | 2765 | 0.6213 | 0.5192 |
0.6514 | 80.0 | 2800 | 0.6170 | 0.5577 |
Framework versions
- Transformers 4.26.1
- Pytorch 2.0.1+cu118
- Datasets 2.12.0
- Tokenizers 0.13.3