t5-end2end-questions-generation
This model is a fine-tuned version of digit82/kolang-t5-base on the korquad_modified_for_t5_qg dataset. It achieves the following results on the evaluation set:
- Loss: 2.1449
Model description
More information needed
Training and evaluation data
KorQuAD V1.0
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0001
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- gradient_accumulation_steps: 16
- total_train_batch_size: 64
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 7
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 3.6685 | 0.66 | 100 | 2.4355 |
| 2.3957 | 1.32 | 200 | 2.2428 |
| 2.1795 | 1.98 | 300 | 2.1664 |
| 1.9408 | 2.65 | 400 | 2.1467 |
| 1.8333 | 3.31 | 500 | 2.1470 |
| 1.7319 | 3.97 | 600 | 2.1194 |
| 1.6095 | 4.63 | 700 | 2.1348 |
| 1.5662 | 5.3 | 800 | 2.1433 |
| 1.5038 | 5.96 | 900 | 2.1319 |
| 1.45 | 6.62 | 1000 | 2.1449 |
Framework versions
- Transformers 4.21.2
- Pytorch 1.12.1+cu113
- Datasets 2.4.0
- Tokenizers 0.12.1
- Downloads last month
- 2