metadata
license: apache-2.0
base_model: bert-base-uncased
tags:
- generated_from_trainer
datasets:
- wikitext
model-index:
- name: run_4
results: []
run_4
This model is a fine-tuned version of bert-base-uncased on the wikitext dataset. It achieves the following results on the evaluation set:
- Loss: 0.2449
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: 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 |
|---|---|---|---|
| 9.1704 | 0.27 | 50 | 8.0057 |
| 7.2118 | 0.55 | 100 | 6.6834 |
| 6.5244 | 0.82 | 150 | 6.3491 |
| 6.2201 | 1.1 | 200 | 6.0229 |
| 5.7189 | 1.37 | 250 | 5.1311 |
| 4.1268 | 1.65 | 300 | 2.9582 |
| 2.4963 | 1.92 | 350 | 1.7429 |
| 1.5611 | 2.2 | 400 | 1.0743 |
| 1.0537 | 2.47 | 450 | 0.7155 |
| 0.7665 | 2.75 | 500 | 0.5189 |
| 0.5947 | 3.02 | 550 | 0.4061 |
| 0.4782 | 3.29 | 600 | 0.3396 |
| 0.4161 | 3.57 | 650 | 0.2976 |
| 0.3785 | 3.84 | 700 | 0.2718 |
| 0.3491 | 4.12 | 750 | 0.2567 |
| 0.3319 | 4.39 | 800 | 0.2488 |
| 0.3286 | 4.67 | 850 | 0.2455 |
| 0.326 | 4.94 | 900 | 0.2449 |
Framework versions
- Transformers 4.33.1
- Pytorch 1.12.1
- Datasets 2.14.6
- Tokenizers 0.13.3