<!-- 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. -->
gpt-expt-sp-v3-K-200-9-mixed
This model is a fine-tuned version of gpt2 on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.0470
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: 0.0005
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- gradient_accumulation_steps: 8
- total_train_batch_size: 256
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 1000
- num_epochs: 500
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss |
---|---|---|---|
0.4566 | 12.75 | 5000 | 0.0648 |
0.0684 | 25.51 | 10000 | 0.0535 |
0.058 | 38.26 | 15000 | 0.0505 |
0.0545 | 51.02 | 20000 | 0.0495 |
0.0527 | 63.77 | 25000 | 0.0491 |
0.0517 | 76.53 | 30000 | 0.0487 |
0.051 | 89.29 | 35000 | 0.0484 |
0.0505 | 102.04 | 40000 | 0.0482 |
0.0502 | 114.79 | 45000 | 0.0480 |
0.0499 | 127.55 | 50000 | 0.0480 |
0.0497 | 140.31 | 55000 | 0.0479 |
0.0495 | 153.06 | 60000 | 0.0478 |
0.0493 | 165.81 | 65000 | 0.0477 |
0.0491 | 178.57 | 70000 | 0.0477 |
0.0489 | 191.33 | 75000 | 0.0476 |
0.0488 | 204.08 | 80000 | 0.0476 |
0.0486 | 216.83 | 85000 | 0.0476 |
0.0485 | 229.59 | 90000 | 0.0475 |
0.0484 | 242.35 | 95000 | 0.0474 |
0.0483 | 255.1 | 100000 | 0.0473 |
0.0482 | 267.86 | 105000 | 0.0473 |
0.0481 | 280.61 | 110000 | 0.0473 |
0.048 | 293.37 | 115000 | 0.0472 |
0.0479 | 306.12 | 120000 | 0.0472 |
0.0478 | 318.88 | 125000 | 0.0472 |
0.0477 | 331.63 | 130000 | 0.0471 |
0.0476 | 344.39 | 135000 | 0.0471 |
0.0475 | 357.14 | 140000 | 0.0471 |
0.0475 | 369.9 | 145000 | 0.0471 |
0.0474 | 382.65 | 150000 | 0.0471 |
0.0473 | 395.41 | 155000 | 0.0470 |
0.0473 | 408.16 | 160000 | 0.0470 |
0.0472 | 420.92 | 165000 | 0.0470 |
0.0472 | 433.67 | 170000 | 0.0470 |
0.0472 | 446.43 | 175000 | 0.0470 |
0.0472 | 459.18 | 180000 | 0.0470 |
0.0471 | 471.94 | 185000 | 0.0470 |
0.0471 | 484.69 | 190000 | 0.0470 |
0.0471 | 497.45 | 195000 | 0.0470 |
Framework versions
- Transformers 4.25.1
- Pytorch 1.13.1+cu117
- Datasets 2.8.0
- Tokenizers 0.13.2