metadata
license: mit
tags:
- generated_from_trainer
datasets:
- generator
model-index:
- name: gpt2-concat-second
results: []
gpt2-concat-second
This model is a fine-tuned version of gpt2 on the generator dataset. It achieves the following results on the evaluation set:
- Loss: 4.4050
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: 64
- eval_batch_size: 64
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 1000
- num_epochs: 10
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 6.7153 | 0.29 | 500 | 5.6161 |
| 5.3436 | 0.58 | 1000 | 5.1899 |
| 4.9993 | 0.87 | 1500 | 4.9392 |
| 4.7264 | 1.16 | 2000 | 4.7817 |
| 4.5632 | 1.45 | 2500 | 4.6599 |
| 4.4515 | 1.74 | 3000 | 4.5490 |
| 4.3483 | 2.02 | 3500 | 4.4674 |
| 4.1412 | 2.31 | 4000 | 4.4283 |
| 4.1268 | 2.6 | 4500 | 4.3805 |
| 4.0932 | 2.89 | 5000 | 4.3336 |
| 3.9281 | 3.18 | 5500 | 4.3330 |
| 3.8693 | 3.47 | 6000 | 4.3021 |
| 3.8701 | 3.76 | 6500 | 4.2746 |
| 3.8108 | 4.05 | 7000 | 4.2753 |
| 3.6096 | 4.34 | 7500 | 4.2838 |
| 3.6425 | 4.63 | 8000 | 4.2588 |
| 3.6484 | 4.92 | 8500 | 4.2325 |
| 3.4344 | 5.21 | 9000 | 4.2856 |
| 3.3896 | 5.49 | 9500 | 4.2764 |
| 3.4182 | 5.78 | 10000 | 4.2599 |
| 3.3427 | 6.07 | 10500 | 4.2920 |
| 3.1434 | 6.36 | 11000 | 4.3128 |
| 3.164 | 6.65 | 11500 | 4.3048 |
| 3.1778 | 6.94 | 12000 | 4.2961 |
| 2.9609 | 7.23 | 12500 | 4.3472 |
| 2.9349 | 7.52 | 13000 | 4.3537 |
| 2.9521 | 7.81 | 13500 | 4.3518 |
| 2.8837 | 8.1 | 14000 | 4.3753 |
| 2.7663 | 8.39 | 14500 | 4.3885 |
| 2.771 | 8.68 | 15000 | 4.3923 |
| 2.7798 | 8.96 | 15500 | 4.3920 |
| 2.6934 | 9.25 | 16000 | 4.4025 |
| 2.685 | 9.54 | 16500 | 4.4043 |
| 2.688 | 9.83 | 17000 | 4.4050 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.11.0+cu113
- Datasets 2.13.0
- Tokenizers 0.13.3