<!-- 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_05-hs_cn
This model is a fine-tuned version of gpt2-medium on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.5395
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 |
---|---|---|---|
73.9845 | 0.02 | 10 | 69.5720 |
46.5054 | 0.04 | 20 | 32.9413 |
13.9547 | 0.06 | 30 | 10.6381 |
7.1893 | 0.08 | 40 | 4.2158 |
4.0234 | 0.1 | 50 | 2.0517 |
1.8675 | 0.12 | 60 | 1.0944 |
1.4112 | 0.14 | 70 | 0.8927 |
1.3181 | 0.16 | 80 | 0.8278 |
1.1966 | 0.18 | 90 | 0.6391 |
1.1266 | 0.2 | 100 | 0.6524 |
1.0109 | 0.22 | 110 | 0.6016 |
1.0176 | 0.24 | 120 | 0.5837 |
1.0522 | 0.26 | 130 | 0.5829 |
1.0986 | 0.28 | 140 | 0.5824 |
1.1256 | 0.3 | 150 | 0.5938 |
1.072 | 0.32 | 160 | 0.5848 |
0.9252 | 0.34 | 170 | 0.5678 |
1.0958 | 0.36 | 180 | 0.5606 |
1.1286 | 0.38 | 190 | 0.5573 |
0.9755 | 0.4 | 200 | 0.5549 |
1.1749 | 0.42 | 210 | 0.5554 |
1.1154 | 0.44 | 220 | 0.5473 |
0.9256 | 0.46 | 230 | 0.5486 |
1.006 | 0.48 | 240 | 0.5464 |
0.9398 | 0.5 | 250 | 0.5485 |
0.9877 | 0.52 | 260 | 0.5478 |
0.9719 | 0.54 | 270 | 0.5442 |
1.0415 | 0.56 | 280 | 0.5441 |
0.9641 | 0.58 | 290 | 0.5449 |
0.8955 | 0.6 | 300 | 0.5462 |
1.0203 | 0.62 | 310 | 0.5415 |
1.1277 | 0.64 | 320 | 0.5415 |
0.943 | 0.66 | 330 | 0.5409 |
0.9258 | 0.68 | 340 | 0.5359 |
0.9217 | 0.7 | 350 | 0.5401 |
0.9456 | 0.72 | 360 | 0.5383 |
1.0309 | 0.74 | 370 | 0.5395 |
Framework versions
- Transformers 4.28.0
- Pytorch 1.11.0+cu113
- Datasets 2.11.0
- Tokenizers 0.12.1