<!-- 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. -->
20230830145026
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: 0.8508
- Accuracy: 0.6176
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: 0.0003
- train_batch_size: 16
- eval_batch_size: 8
- seed: 11
- 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 | 340 | 0.7278 | 0.5219 |
0.7489 | 2.0 | 680 | 0.7277 | 0.5 |
0.7435 | 3.0 | 1020 | 0.7386 | 0.5 |
0.7435 | 4.0 | 1360 | 0.7302 | 0.5 |
0.7399 | 5.0 | 1700 | 0.7256 | 0.5 |
0.7423 | 6.0 | 2040 | 0.7256 | 0.5 |
0.7423 | 7.0 | 2380 | 0.7297 | 0.5 |
0.7409 | 8.0 | 2720 | 0.7365 | 0.5 |
0.7408 | 9.0 | 3060 | 0.7305 | 0.5 |
0.7408 | 10.0 | 3400 | 0.7599 | 0.5 |
0.7394 | 11.0 | 3740 | 0.7301 | 0.5 |
0.7356 | 12.0 | 4080 | 0.7531 | 0.5 |
0.7356 | 13.0 | 4420 | 0.7362 | 0.5 |
0.7351 | 14.0 | 4760 | 0.7306 | 0.5 |
0.7383 | 15.0 | 5100 | 0.7366 | 0.5 |
0.7383 | 16.0 | 5440 | 0.7526 | 0.5 |
0.7351 | 17.0 | 5780 | 0.7400 | 0.5 |
0.7354 | 18.0 | 6120 | 0.7302 | 0.5 |
0.7354 | 19.0 | 6460 | 0.7251 | 0.5 |
0.7328 | 20.0 | 6800 | 0.7274 | 0.5 |
0.733 | 21.0 | 7140 | 0.7263 | 0.5 |
0.733 | 22.0 | 7480 | 0.7423 | 0.5 |
0.7338 | 23.0 | 7820 | 0.7297 | 0.5 |
0.733 | 24.0 | 8160 | 0.7239 | 0.5 |
0.733 | 25.0 | 8500 | 0.7245 | 0.5 |
0.733 | 26.0 | 8840 | 0.7267 | 0.5 |
0.7332 | 27.0 | 9180 | 0.7595 | 0.5 |
0.7327 | 28.0 | 9520 | 0.7233 | 0.5 |
0.7327 | 29.0 | 9860 | 0.7339 | 0.5 |
0.7319 | 30.0 | 10200 | 0.7193 | 0.5047 |
0.7283 | 31.0 | 10540 | 0.6910 | 0.5987 |
0.7283 | 32.0 | 10880 | 0.7071 | 0.6614 |
0.7085 | 33.0 | 11220 | 0.6750 | 0.6489 |
0.6778 | 34.0 | 11560 | 0.6710 | 0.6614 |
0.6778 | 35.0 | 11900 | 0.7262 | 0.6426 |
0.6531 | 36.0 | 12240 | 0.6715 | 0.6677 |
0.6288 | 37.0 | 12580 | 0.7028 | 0.6536 |
0.6288 | 38.0 | 12920 | 0.7200 | 0.6505 |
0.6058 | 39.0 | 13260 | 0.7034 | 0.6536 |
0.5953 | 40.0 | 13600 | 0.6975 | 0.6458 |
0.5953 | 41.0 | 13940 | 0.6987 | 0.6489 |
0.5723 | 42.0 | 14280 | 0.7472 | 0.6426 |
0.555 | 43.0 | 14620 | 0.7437 | 0.6426 |
0.555 | 44.0 | 14960 | 0.7673 | 0.6066 |
0.5439 | 45.0 | 15300 | 0.7111 | 0.6348 |
0.533 | 46.0 | 15640 | 0.7657 | 0.6270 |
0.533 | 47.0 | 15980 | 0.8004 | 0.6332 |
0.525 | 48.0 | 16320 | 0.7423 | 0.6144 |
0.5098 | 49.0 | 16660 | 0.7131 | 0.6505 |
0.5078 | 50.0 | 17000 | 0.7751 | 0.6144 |
0.5078 | 51.0 | 17340 | 0.7272 | 0.6505 |
0.4987 | 52.0 | 17680 | 0.7666 | 0.6473 |
0.4892 | 53.0 | 18020 | 0.8000 | 0.6379 |
0.4892 | 54.0 | 18360 | 0.7950 | 0.6348 |
0.4813 | 55.0 | 18700 | 0.7802 | 0.6317 |
0.4756 | 56.0 | 19040 | 0.8083 | 0.6129 |
0.4756 | 57.0 | 19380 | 0.7916 | 0.6364 |
0.4685 | 58.0 | 19720 | 0.8475 | 0.6066 |
0.4694 | 59.0 | 20060 | 0.7576 | 0.6646 |
0.4694 | 60.0 | 20400 | 0.8712 | 0.6113 |
0.4513 | 61.0 | 20740 | 0.8164 | 0.6160 |
0.4492 | 62.0 | 21080 | 0.8174 | 0.6270 |
0.4492 | 63.0 | 21420 | 0.8198 | 0.6270 |
0.4458 | 64.0 | 21760 | 0.8006 | 0.6395 |
0.4351 | 65.0 | 22100 | 0.7962 | 0.6520 |
0.4351 | 66.0 | 22440 | 0.7911 | 0.6301 |
0.4346 | 67.0 | 22780 | 0.8150 | 0.6395 |
0.4356 | 68.0 | 23120 | 0.8336 | 0.6301 |
0.4356 | 69.0 | 23460 | 0.8777 | 0.6097 |
0.4283 | 70.0 | 23800 | 0.8313 | 0.6191 |
0.4212 | 71.0 | 24140 | 0.8191 | 0.6034 |
0.4212 | 72.0 | 24480 | 0.8384 | 0.6113 |
0.4197 | 73.0 | 24820 | 0.8532 | 0.6176 |
0.4172 | 74.0 | 25160 | 0.8398 | 0.6097 |
0.4112 | 75.0 | 25500 | 0.8536 | 0.6113 |
0.4112 | 76.0 | 25840 | 0.8130 | 0.6176 |
0.4104 | 77.0 | 26180 | 0.8463 | 0.6082 |
0.4128 | 78.0 | 26520 | 0.8342 | 0.6160 |
0.4128 | 79.0 | 26860 | 0.8502 | 0.6176 |
0.4004 | 80.0 | 27200 | 0.8508 | 0.6176 |
Framework versions
- Transformers 4.26.1
- Pytorch 2.0.1+cu118
- Datasets 2.12.0
- Tokenizers 0.13.3