|
--- |
|
license: mit |
|
tags: |
|
- generated_from_trainer |
|
datasets: |
|
- generator |
|
model-index: |
|
- name: gpt2_left_out_switchboard |
|
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_left_out_switchboard |
|
|
|
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: 3.9378 |
|
|
|
## 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 | |
|
|:-------------:|:-----:|:-----:|:---------------:| |
|
| 5.983 | 0.24 | 500 | 5.0786 | |
|
| 4.7603 | 0.48 | 1000 | 4.6865 | |
|
| 4.4521 | 0.73 | 1500 | 4.4635 | |
|
| 4.2512 | 0.97 | 2000 | 4.3124 | |
|
| 4.0458 | 1.21 | 2500 | 4.2272 | |
|
| 3.9687 | 1.45 | 3000 | 4.1443 | |
|
| 3.9024 | 1.69 | 3500 | 4.0705 | |
|
| 3.8439 | 1.93 | 4000 | 4.0057 | |
|
| 3.6791 | 2.18 | 4500 | 3.9845 | |
|
| 3.6259 | 2.42 | 5000 | 3.9471 | |
|
| 3.6137 | 2.66 | 5500 | 3.9057 | |
|
| 3.592 | 2.9 | 6000 | 3.8654 | |
|
| 3.4438 | 3.14 | 6500 | 3.8758 | |
|
| 3.3844 | 3.38 | 7000 | 3.8570 | |
|
| 3.3977 | 3.63 | 7500 | 3.8324 | |
|
| 3.4015 | 3.87 | 8000 | 3.8053 | |
|
| 3.2638 | 4.11 | 8500 | 3.8300 | |
|
| 3.1771 | 4.35 | 9000 | 3.8250 | |
|
| 3.1914 | 4.59 | 9500 | 3.8070 | |
|
| 3.1993 | 4.84 | 10000 | 3.7853 | |
|
| 3.1089 | 5.08 | 10500 | 3.8146 | |
|
| 2.9539 | 5.32 | 11000 | 3.8262 | |
|
| 2.9853 | 5.56 | 11500 | 3.8173 | |
|
| 2.9984 | 5.8 | 12000 | 3.8020 | |
|
| 2.9462 | 6.04 | 12500 | 3.8259 | |
|
| 2.7343 | 6.29 | 13000 | 3.8527 | |
|
| 2.7724 | 6.53 | 13500 | 3.8499 | |
|
| 2.7817 | 6.77 | 14000 | 3.8423 | |
|
| 2.7789 | 7.01 | 14500 | 3.8510 | |
|
| 2.5477 | 7.25 | 15000 | 3.8873 | |
|
| 2.5643 | 7.5 | 15500 | 3.8904 | |
|
| 2.5842 | 7.74 | 16000 | 3.8896 | |
|
| 2.5913 | 7.98 | 16500 | 3.8858 | |
|
| 2.4293 | 8.22 | 17000 | 3.9177 | |
|
| 2.4253 | 8.46 | 17500 | 3.9231 | |
|
| 2.4274 | 8.7 | 18000 | 3.9240 | |
|
| 2.4331 | 8.95 | 18500 | 3.9254 | |
|
| 2.362 | 9.19 | 19000 | 3.9346 | |
|
| 2.3519 | 9.43 | 19500 | 3.9373 | |
|
| 2.3498 | 9.67 | 20000 | 3.9378 | |
|
| 2.3461 | 9.91 | 20500 | 3.9378 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.26.1 |
|
- Pytorch 1.11.0+cu113 |
|
- Datasets 2.13.0 |
|
- Tokenizers 0.13.3 |
|
|