polynomial_1750_7e-4_new
This model is a fine-tuned version of gpt2 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 2.8314
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: 0.0007
- 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: polynomial
- lr_scheduler_warmup_steps: 250
- training_steps: 1750
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss |
---|---|---|---|
9.0487 | 0.2058 | 50 | 7.2380 |
6.6233 | 0.4117 | 100 | 6.0822 |
5.8244 | 0.6175 | 150 | 5.5096 |
5.348 | 0.8234 | 200 | 5.0940 |
4.9539 | 1.0292 | 250 | 4.7446 |
4.5822 | 1.2351 | 300 | 4.4289 |
4.3064 | 1.4409 | 350 | 4.1720 |
4.057 | 1.6468 | 400 | 3.8873 |
3.803 | 1.8526 | 450 | 3.6405 |
3.5656 | 2.0585 | 500 | 3.4743 |
3.3722 | 2.2643 | 550 | 3.3687 |
3.3005 | 2.4702 | 600 | 3.2950 |
3.2308 | 2.6760 | 650 | 3.2288 |
3.1753 | 2.8818 | 700 | 3.1707 |
3.0716 | 3.0877 | 750 | 3.1226 |
2.9645 | 3.2935 | 800 | 3.0926 |
2.9294 | 3.4994 | 850 | 3.0588 |
2.9324 | 3.7052 | 900 | 3.0269 |
2.9058 | 3.9111 | 950 | 2.9988 |
2.7861 | 4.1169 | 1000 | 2.9782 |
2.7313 | 4.3228 | 1050 | 2.9634 |
2.7262 | 4.5286 | 1100 | 2.9476 |
2.717 | 4.7345 | 1150 | 2.9242 |
2.7049 | 4.9403 | 1200 | 2.9056 |
2.5917 | 5.1462 | 1250 | 2.8997 |
2.5555 | 5.3520 | 1300 | 2.8901 |
2.5438 | 5.5578 | 1350 | 2.8790 |
2.5467 | 5.7637 | 1400 | 2.8656 |
2.5394 | 5.9695 | 1450 | 2.8530 |
2.4139 | 6.1754 | 1500 | 2.8548 |
2.4083 | 6.3812 | 1550 | 2.8490 |
2.4087 | 6.5871 | 1600 | 2.8421 |
2.3971 | 6.7929 | 1650 | 2.8348 |
2.3994 | 6.9988 | 1700 | 2.8291 |
2.3092 | 7.2046 | 1750 | 2.8314 |
Framework versions
- Transformers 4.40.1
- Pytorch 2.3.0+cu121
- Datasets 2.19.0
- Tokenizers 0.19.1
- Downloads last month
- 3