|
|
--- |
|
|
library_name: transformers |
|
|
tags: |
|
|
- generated_from_trainer |
|
|
metrics: |
|
|
- accuracy |
|
|
- bleu |
|
|
model-index: |
|
|
- name: duo-predict-gpt2-medium-wikitext |
|
|
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. --> |
|
|
|
|
|
# duo-predict-gpt2-medium-wikitext |
|
|
|
|
|
This model is a fine-tuned version of [](https://huggingface.co/) on an unknown dataset. |
|
|
It achieves the following results on the evaluation set: |
|
|
- Loss: 2.2546 |
|
|
- Accuracy: 0.0073 |
|
|
- Perplexity: 9.5311 |
|
|
- Bleu: 1.0 |
|
|
|
|
|
## 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: 0.0001 |
|
|
- train_batch_size: 64 |
|
|
- eval_batch_size: 64 |
|
|
- seed: 42 |
|
|
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments |
|
|
- lr_scheduler_type: linear |
|
|
- lr_scheduler_warmup_ratio: 0.1 |
|
|
- num_epochs: 5 |
|
|
|
|
|
### Training results |
|
|
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Perplexity | Bleu | |
|
|
|:-------------:|:------:|:-----:|:---------------:|:--------:|:----------:|:----:| |
|
|
| 7.6654 | 0.1403 | 500 | 3.7315 | 0.0073 | 41.7396 | 1.0 | |
|
|
| 7.0276 | 0.2807 | 1000 | 3.4735 | 0.0073 | 32.2490 | 1.0 | |
|
|
| 6.4629 | 0.4210 | 1500 | 3.1863 | 0.0073 | 24.1987 | 1.0 | |
|
|
| 5.9671 | 0.5613 | 2000 | 2.9542 | 0.0073 | 19.1873 | 1.0 | |
|
|
| 5.6969 | 0.7017 | 2500 | 2.8233 | 0.0073 | 16.8331 | 1.0 | |
|
|
| 5.5077 | 0.8420 | 3000 | 2.7351 | 0.0073 | 15.4112 | 1.0 | |
|
|
| 5.3536 | 0.9823 | 3500 | 2.6607 | 0.0073 | 14.3059 | 1.0 | |
|
|
| 5.2099 | 1.1226 | 4000 | 2.6000 | 0.0073 | 13.4641 | 1.0 | |
|
|
| 5.1158 | 1.2630 | 4500 | 2.5493 | 0.0073 | 12.7980 | 1.0 | |
|
|
| 5.0453 | 1.4033 | 5000 | 2.5125 | 0.0073 | 12.3362 | 1.0 | |
|
|
| 4.955 | 1.5436 | 5500 | 2.4806 | 0.0073 | 11.9489 | 1.0 | |
|
|
| 4.9157 | 1.6840 | 6000 | 2.4537 | 0.0073 | 11.6310 | 1.0 | |
|
|
| 4.8756 | 1.8243 | 6500 | 2.4300 | 0.0073 | 11.3584 | 1.0 | |
|
|
| 4.844 | 1.9646 | 7000 | 2.4100 | 0.0073 | 11.1342 | 1.0 | |
|
|
| 4.7136 | 2.1050 | 7500 | 2.3948 | 0.0073 | 10.9657 | 1.0 | |
|
|
| 4.6911 | 2.2453 | 8000 | 2.3805 | 0.0073 | 10.8105 | 1.0 | |
|
|
| 4.6741 | 2.3856 | 8500 | 2.3668 | 0.0073 | 10.6637 | 1.0 | |
|
|
| 4.6485 | 2.5260 | 9000 | 2.3538 | 0.0073 | 10.5257 | 1.0 | |
|
|
| 4.623 | 2.6663 | 9500 | 2.3416 | 0.0073 | 10.3976 | 1.0 | |
|
|
| 4.6016 | 2.8066 | 10000 | 2.3303 | 0.0073 | 10.2806 | 1.0 | |
|
|
| 4.5823 | 2.9470 | 10500 | 2.3202 | 0.0073 | 10.1776 | 1.0 | |
|
|
| 4.4802 | 3.0873 | 11000 | 2.3143 | 0.0073 | 10.1182 | 1.0 | |
|
|
| 4.4671 | 3.2276 | 11500 | 2.3073 | 0.0073 | 10.0469 | 1.0 | |
|
|
| 4.4557 | 3.3679 | 12000 | 2.3006 | 0.0073 | 9.9800 | 1.0 | |
|
|
| 4.4437 | 3.5083 | 12500 | 2.2928 | 0.0073 | 9.9023 | 1.0 | |
|
|
| 4.4402 | 3.6486 | 13000 | 2.2862 | 0.0073 | 9.8375 | 1.0 | |
|
|
| 4.4482 | 3.7889 | 13500 | 2.2800 | 0.0073 | 9.7763 | 1.0 | |
|
|
| 4.4279 | 3.9293 | 14000 | 2.2752 | 0.0073 | 9.7303 | 1.0 | |
|
|
| 4.3188 | 4.0696 | 14500 | 2.2730 | 0.0073 | 9.7087 | 1.0 | |
|
|
| 4.3193 | 4.2099 | 15000 | 2.2691 | 0.0073 | 9.6704 | 1.0 | |
|
|
| 4.3158 | 4.3503 | 15500 | 2.2652 | 0.0073 | 9.6329 | 1.0 | |
|
|
| 4.3196 | 4.4906 | 16000 | 2.2619 | 0.0073 | 9.6012 | 1.0 | |
|
|
| 4.2946 | 4.6309 | 16500 | 2.2589 | 0.0073 | 9.5722 | 1.0 | |
|
|
| 4.3078 | 4.7713 | 17000 | 2.2564 | 0.0073 | 9.5487 | 1.0 | |
|
|
| 4.2974 | 4.9116 | 17500 | 2.2546 | 0.0073 | 9.5311 | 1.0 | |
|
|
|
|
|
|
|
|
### Framework versions |
|
|
|
|
|
- Transformers 4.49.0 |
|
|
- Pytorch 2.6.0+cu124 |
|
|
- Datasets 3.3.2 |
|
|
- Tokenizers 0.21.0 |
|
|
|
