|
|
--- |
|
|
license: mit |
|
|
tags: |
|
|
- generated_from_trainer |
|
|
datasets: |
|
|
- generator |
|
|
model-index: |
|
|
- name: gpt2-dp-3 |
|
|
results: [] |
|
|
--- |
|
|
|
|
|
<!-- 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-dp-3 |
|
|
|
|
|
This model is a fine-tuned version of [gpt2](https://huggingface.co/gpt2) on the generator dataset. |
|
|
It achieves the following results on the evaluation set: |
|
|
- Loss: 4.4076 |
|
|
|
|
|
## 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 | |
|
|
|:-------------:|:-----:|:-----:|:---------------:| |
|
|
| 6.7156 | 0.27 | 500 | 5.6535 | |
|
|
| 5.3578 | 0.53 | 1000 | 5.2045 | |
|
|
| 5.0077 | 0.8 | 1500 | 4.9659 | |
|
|
| 4.7593 | 1.07 | 2000 | 4.8126 | |
|
|
| 4.5687 | 1.34 | 2500 | 4.7029 | |
|
|
| 4.4766 | 1.6 | 3000 | 4.5953 | |
|
|
| 4.3917 | 1.87 | 3500 | 4.5056 | |
|
|
| 4.2228 | 2.14 | 4000 | 4.4626 | |
|
|
| 4.1279 | 2.4 | 4500 | 4.4147 | |
|
|
| 4.1019 | 2.67 | 5000 | 4.3627 | |
|
|
| 4.0683 | 2.94 | 5500 | 4.3206 | |
|
|
| 3.869 | 3.21 | 6000 | 4.3295 | |
|
|
| 3.8494 | 3.47 | 6500 | 4.3034 | |
|
|
| 3.8533 | 3.74 | 7000 | 4.2734 | |
|
|
| 3.8342 | 4.01 | 7500 | 4.2661 | |
|
|
| 3.5799 | 4.27 | 8000 | 4.2817 | |
|
|
| 3.6163 | 4.54 | 8500 | 4.2654 | |
|
|
| 3.6245 | 4.81 | 9000 | 4.2402 | |
|
|
| 3.5328 | 5.07 | 9500 | 4.2692 | |
|
|
| 3.3455 | 5.34 | 10000 | 4.2804 | |
|
|
| 3.3898 | 5.61 | 10500 | 4.2662 | |
|
|
| 3.3933 | 5.88 | 11000 | 4.2519 | |
|
|
| 3.2239 | 6.14 | 11500 | 4.3025 | |
|
|
| 3.1152 | 6.41 | 12000 | 4.3098 | |
|
|
| 3.14 | 6.68 | 12500 | 4.3060 | |
|
|
| 3.1585 | 6.94 | 13000 | 4.2908 | |
|
|
| 2.9392 | 7.21 | 13500 | 4.3478 | |
|
|
| 2.9031 | 7.48 | 14000 | 4.3549 | |
|
|
| 2.9201 | 7.75 | 14500 | 4.3523 | |
|
|
| 2.9044 | 8.01 | 15000 | 4.3650 | |
|
|
| 2.7244 | 8.28 | 15500 | 4.3877 | |
|
|
| 2.7371 | 8.55 | 16000 | 4.3929 | |
|
|
| 2.745 | 8.81 | 16500 | 4.3943 | |
|
|
| 2.7233 | 9.08 | 17000 | 4.4028 | |
|
|
| 2.6481 | 9.35 | 17500 | 4.4060 | |
|
|
| 2.6578 | 9.62 | 18000 | 4.4077 | |
|
|
| 2.6554 | 9.88 | 18500 | 4.4076 | |
|
|
|
|
|
|
|
|
### Framework versions |
|
|
|
|
|
- Transformers 4.26.1 |
|
|
- Pytorch 1.11.0+cu113 |
|
|
- Datasets 2.13.0 |
|
|
- Tokenizers 0.13.3 |
|
|
|
