<!-- 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. -->
gpt2-kl_1_03-hs_cn_decay
This model is a fine-tuned version of gpt2-medium on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.5373
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: 8
- eval_batch_size: 4
- seed: 21
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 3.0
Training results
Training Loss | Epoch | Step | Validation Loss |
---|---|---|---|
73.692 | 0.02 | 10 | 69.5606 |
46.2212 | 0.04 | 20 | 32.9081 |
13.6702 | 0.06 | 30 | 10.6389 |
5.9581 | 0.08 | 40 | 4.3665 |
2.8384 | 0.1 | 50 | 1.9575 |
1.4754 | 0.12 | 60 | 1.0434 |
1.1362 | 0.14 | 70 | 0.8548 |
0.9905 | 0.16 | 80 | 0.7224 |
0.8906 | 0.18 | 90 | 0.6321 |
0.8441 | 0.2 | 100 | 0.6041 |
0.7437 | 0.22 | 110 | 0.5988 |
0.7504 | 0.24 | 120 | 0.5891 |
0.7622 | 0.26 | 130 | 0.6082 |
0.7901 | 0.28 | 140 | 0.5893 |
0.7747 | 0.3 | 150 | 0.5730 |
0.7417 | 0.32 | 160 | 0.5650 |
0.649 | 0.34 | 170 | 0.5639 |
0.7707 | 0.36 | 180 | 0.5638 |
0.7895 | 0.38 | 190 | 0.5581 |
0.6949 | 0.4 | 200 | 0.5564 |
0.8323 | 0.42 | 210 | 0.5543 |
0.784 | 0.44 | 220 | 0.5509 |
0.6406 | 0.46 | 230 | 0.5526 |
0.6926 | 0.48 | 240 | 0.5498 |
0.664 | 0.5 | 250 | 0.5520 |
0.6776 | 0.52 | 260 | 0.5505 |
0.6794 | 0.54 | 270 | 0.5455 |
0.7237 | 0.56 | 280 | 0.5465 |
0.6952 | 0.58 | 290 | 0.5449 |
0.6384 | 0.6 | 300 | 0.5452 |
0.6794 | 0.62 | 310 | 0.5411 |
0.7838 | 0.64 | 320 | 0.5431 |
0.6222 | 0.66 | 330 | 0.5387 |
0.6246 | 0.68 | 340 | 0.5363 |
0.6279 | 0.7 | 350 | 0.5377 |
0.628 | 0.72 | 360 | 0.5366 |
0.7132 | 0.74 | 370 | 0.5373 |
Framework versions
- Transformers 4.28.0
- Pytorch 1.11.0+cu113
- Datasets 2.11.0
- Tokenizers 0.12.1