pretraining6
This model is a fine-tuned version of gpt2 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 2.9244
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: 0.0006
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- gradient_accumulation_steps: 10
- total_train_batch_size: 320
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: constant
- lr_scheduler_warmup_steps: 250
- training_steps: 2500
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 7.5942 | 0.1830 | 50 | 6.9837 |
| 6.7139 | 0.3660 | 100 | 6.4890 |
| 6.3586 | 0.5490 | 150 | 6.1468 |
| 6.0068 | 0.7321 | 200 | 5.8162 |
| 5.7156 | 0.9151 | 250 | 5.5687 |
| 5.4902 | 1.0981 | 300 | 5.3574 |
| 5.2778 | 1.2811 | 350 | 5.1752 |
| 5.078 | 1.4641 | 400 | 4.9877 |
| 4.905 | 1.6471 | 450 | 4.8093 |
| 4.7396 | 1.8302 | 500 | 4.6300 |
| 4.5488 | 2.0132 | 550 | 4.4533 |
| 4.2909 | 2.1962 | 600 | 4.2386 |
| 4.1235 | 2.3792 | 650 | 3.9890 |
| 3.9081 | 2.5622 | 700 | 3.7933 |
| 3.7373 | 2.7452 | 750 | 3.6421 |
| 3.6011 | 2.9283 | 800 | 3.5265 |
| 3.4526 | 3.1113 | 850 | 3.4465 |
| 3.3523 | 3.2943 | 900 | 3.3867 |
| 3.2917 | 3.4773 | 950 | 3.3297 |
| 3.2536 | 3.6603 | 1000 | 3.2808 |
| 3.2277 | 3.8433 | 1050 | 3.2435 |
| 3.1699 | 4.0264 | 1100 | 3.1971 |
| 3.0158 | 4.2094 | 1150 | 3.1710 |
| 3.0104 | 4.3924 | 1200 | 3.1499 |
| 2.9946 | 4.5754 | 1250 | 3.1194 |
| 2.9814 | 4.7584 | 1300 | 3.0988 |
| 2.9686 | 4.9414 | 1350 | 3.0700 |
| 2.8425 | 5.1245 | 1400 | 3.0559 |
| 2.8039 | 5.3075 | 1450 | 3.0437 |
| 2.8121 | 5.4905 | 1500 | 3.0285 |
| 2.8078 | 5.6735 | 1550 | 3.0128 |
| 2.7996 | 5.8565 | 1600 | 2.9962 |
| 2.7607 | 6.0395 | 1650 | 2.9871 |
| 2.6212 | 6.2225 | 1700 | 2.9845 |
| 2.6638 | 6.4056 | 1750 | 2.9746 |
| 2.6603 | 6.5886 | 1800 | 2.9660 |
| 2.6674 | 6.7716 | 1850 | 2.9510 |
| 2.6741 | 6.9546 | 1900 | 2.9379 |
| 2.5313 | 7.1376 | 1950 | 2.9474 |
| 2.5107 | 7.3206 | 2000 | 2.9465 |
| 2.5358 | 7.5037 | 2050 | 2.9403 |
| 2.5552 | 7.6867 | 2100 | 2.9303 |
| 2.5691 | 7.8697 | 2150 | 2.9200 |
| 2.5008 | 8.0527 | 2200 | 2.9241 |
| 2.3855 | 8.2357 | 2250 | 2.9314 |
| 2.4215 | 8.4187 | 2300 | 2.9285 |
| 2.4488 | 8.6018 | 2350 | 2.9217 |
| 2.46 | 8.7848 | 2400 | 2.9110 |
| 2.468 | 8.9678 | 2450 | 2.9044 |
| 2.3004 | 9.1508 | 2500 | 2.9244 |
Framework versions
- Transformers 4.40.1
- Pytorch 2.3.0+cu121
- Datasets 2.19.0
- Tokenizers 0.19.1
- Downloads last month
- 3