|
|
--- |
|
|
license: apache-2.0 |
|
|
base_model: bert-base-uncased |
|
|
tags: |
|
|
- generated_from_trainer |
|
|
datasets: |
|
|
- wikitext |
|
|
model-index: |
|
|
- name: run_2 |
|
|
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. --> |
|
|
|
|
|
# run_2 |
|
|
|
|
|
This model is a fine-tuned version of [bert-base-uncased](https://huggingface.co/bert-base-uncased) on the wikitext dataset. |
|
|
It achieves the following results on the evaluation set: |
|
|
- Loss: 0.9502 |
|
|
|
|
|
## 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: 100 |
|
|
- num_epochs: 5 |
|
|
|
|
|
### Training results |
|
|
|
|
|
| Training Loss | Epoch | Step | Validation Loss | |
|
|
|:-------------:|:-----:|:----:|:---------------:| |
|
|
| 8.4559 | 0.27 | 50 | 7.1236 | |
|
|
| 6.8523 | 0.55 | 100 | 6.6676 | |
|
|
| 6.6103 | 0.82 | 150 | 6.5582 | |
|
|
| 6.2417 | 1.1 | 200 | 5.6994 | |
|
|
| 4.9738 | 1.37 | 250 | 4.3440 | |
|
|
| 4.1043 | 1.65 | 300 | 3.7804 | |
|
|
| 3.4265 | 1.92 | 350 | 3.0136 | |
|
|
| 2.7667 | 2.2 | 400 | 2.5318 | |
|
|
| 2.3538 | 2.47 | 450 | 2.0903 | |
|
|
| 1.9591 | 2.75 | 500 | 1.7367 | |
|
|
| 1.6652 | 3.02 | 550 | 1.5016 | |
|
|
| 1.4318 | 3.29 | 600 | 1.3162 | |
|
|
| 1.275 | 3.57 | 650 | 1.1657 | |
|
|
| 1.1553 | 3.84 | 700 | 1.0655 | |
|
|
| 1.0629 | 4.12 | 750 | 1.0029 | |
|
|
| 1.0029 | 4.39 | 800 | 0.9683 | |
|
|
| 0.9881 | 4.67 | 850 | 0.9536 | |
|
|
| 0.9779 | 4.94 | 900 | 0.9502 | |
|
|
|
|
|
|
|
|
### Framework versions |
|
|
|
|
|
- Transformers 4.33.1 |
|
|
- Pytorch 1.12.1 |
|
|
- Datasets 2.14.6 |
|
|
- Tokenizers 0.13.3 |
|
|
|
