| --- |
| library_name: transformers |
| license: mit |
| base_model: facebook/mbart-large-50 |
| tags: |
| - generated_from_trainer |
| model-index: |
| - name: scenegraph_image_captioning_vi |
| 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. --> |
|
|
| # scenegraph_image_captioning_vi |
| |
| This model is a fine-tuned version of [facebook/mbart-large-50](https://huggingface.co/facebook/mbart-large-50) on an unknown dataset. |
| It achieves the following results on the evaluation set: |
| - Loss: 3.1815 |
| - Bleu4: 16.1492 |
| |
| ## 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: 3e-05 |
| - train_batch_size: 48 |
| - eval_batch_size: 48 |
| - seed: 42 |
| - gradient_accumulation_steps: 2 |
| - total_train_batch_size: 96 |
| - 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_steps: 350 |
| - num_epochs: 30 |
| |
| ### Training results |
| |
| | Training Loss | Epoch | Step | Bleu4 | Validation Loss | |
| |:-------------:|:-----:|:-----:|:-------:|:---------------:| |
| | 3.5137 | 1.0 | 3083 | 11.8025 | 3.4747 | |
| | 3.3145 | 2.0 | 6166 | 13.309 | 3.3252 | |
| | 3.1957 | 3.0 | 9249 | 14.2144 | 3.2632 | |
| | 3.1449 | 4.0 | 12332 | 14.8718 | 3.2247 | |
| | 3.1078 | 5.0 | 15415 | 15.1815 | 3.1963 | |
| | 3.0352 | 6.0 | 18498 | 15.2137 | 3.1796 | |
| | 3.0147 | 7.0 | 21581 | 15.7702 | 3.1697 | |
| | 2.971 | 8.0 | 24664 | 16.0201 | 3.1662 | |
| | 2.9144 | 9.0 | 27747 | 16.1495 | 3.1623 | |
| | 2.9037 | 10.0 | 30830 | 16.0747 | 3.1611 | |
| | 2.9178 | 11.0 | 33913 | 16.2077 | 3.1575 | |
| | 2.8687 | 12.0 | 36996 | 16.1345 | 3.1594 | |
| | 2.8426 | 13.0 | 40079 | 16.2958 | 3.1618 | |
| | 2.8058 | 14.0 | 43162 | 16.2391 | 3.1639 | |
| | 2.814 | 15.0 | 46245 | 16.341 | 3.1722 | |
| | 2.782 | 16.0 | 49328 | 3.1683 | 16.0129 | |
| | 2.7853 | 17.0 | 52411 | 3.1772 | 16.1951 | |
| | 2.7313 | 18.0 | 55494 | 3.1846 | 16.0554 | |
| | 2.7525 | 19.0 | 58577 | 3.1876 | 16.0271 | |
| | 2.7183 | 20.0 | 61660 | 3.1815 | 16.1492 | |
| |
| |
| ### Framework versions |
| |
| - Transformers 4.51.0 |
| - Pytorch 2.11.0+cu128 |
| - Datasets 4.0.0 |
| - Tokenizers 0.21.4 |
| |
