metadata
license: apache-2.0
base_model: studio-ousia/luke-base
tags:
- generated_from_trainer
datasets:
- conll2003
model-index:
- name: luke-base_finetuned_conll2003
results: []
luke-base_finetuned_conll2003
This model is a fine-tuned version of studio-ousia/luke-base on the conll2003 dataset. It achieves the following results on the evaluation set:
- Loss: 0.0959
- F1-type-match: 0.7530
- F1-partial: 0.7885
- F1-strict: 0.7477
- F1-exact: 0.7828
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: 16
- eval_batch_size: 16
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 64
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 5
Training results
| Training Loss | Epoch | Step | Validation Loss | F1-type-match | F1-partial | F1-strict | F1-exact |
|---|---|---|---|---|---|---|---|
| No log | 1.0 | 219 | 0.0796 | 0.3002 | 0.3178 | 0.2969 | 0.3146 |
| 0.1554 | 2.0 | 439 | 0.0724 | 0.3889 | 0.4078 | 0.3859 | 0.4048 |
| 0.0256 | 3.0 | 658 | 0.0832 | 0.5340 | 0.5610 | 0.5300 | 0.5568 |
| 0.0124 | 4.0 | 878 | 0.0918 | 0.6875 | 0.7207 | 0.6826 | 0.7156 |
| 0.0062 | 4.99 | 1095 | 0.0959 | 0.7530 | 0.7885 | 0.7477 | 0.7828 |
Framework versions
- Transformers 4.35.0
- Pytorch 2.0.0
- Datasets 2.1.0
- Tokenizers 0.14.1