|
--- |
|
license: mit |
|
base_model: microsoft/deberta-v3-base |
|
tags: |
|
- generated_from_trainer |
|
model-index: |
|
- name: deberta-v3-base-zalo-v1 |
|
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. --> |
|
|
|
# deberta-v3-base-zalo-v1 |
|
|
|
This model is a fine-tuned version of [microsoft/deberta-v3-base](https://huggingface.co/microsoft/deberta-v3-base) on the None dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 1.3848 |
|
- Map@3: 0.6788 |
|
|
|
## 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: 2e-05 |
|
- train_batch_size: 8 |
|
- eval_batch_size: 2 |
|
- seed: 42 |
|
- gradient_accumulation_steps: 8 |
|
- total_train_batch_size: 64 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: cosine |
|
- lr_scheduler_warmup_ratio: 0.1 |
|
- num_epochs: 1 |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Map@3 | |
|
|:-------------:|:-----:|:----:|:---------------:|:------:| |
|
| 1.3875 | 0.02 | 25 | 1.3864 | 0.4245 | |
|
| 1.388 | 0.04 | 50 | 1.3863 | 0.4449 | |
|
| 1.388 | 0.06 | 75 | 1.3863 | 0.4549 | |
|
| 1.3862 | 0.08 | 100 | 1.3863 | 0.4674 | |
|
| 1.3867 | 0.1 | 125 | 1.3862 | 0.5061 | |
|
| 1.3859 | 0.13 | 150 | 1.3861 | 0.5347 | |
|
| 1.3871 | 0.15 | 175 | 1.3860 | 0.5556 | |
|
| 1.3851 | 0.17 | 200 | 1.3860 | 0.5764 | |
|
| 1.3865 | 0.19 | 225 | 1.3859 | 0.5751 | |
|
| 1.3853 | 0.21 | 250 | 1.3859 | 0.5864 | |
|
| 1.3863 | 0.23 | 275 | 1.3858 | 0.6128 | |
|
| 1.3858 | 0.25 | 300 | 1.3857 | 0.6176 | |
|
| 1.3875 | 0.27 | 325 | 1.3857 | 0.6350 | |
|
| 1.3863 | 0.29 | 350 | 1.3856 | 0.6432 | |
|
| 1.3858 | 0.31 | 375 | 1.3856 | 0.6502 | |
|
| 1.3861 | 0.34 | 400 | 1.3855 | 0.6589 | |
|
| 1.3867 | 0.36 | 425 | 1.3855 | 0.6593 | |
|
| 1.387 | 0.38 | 450 | 1.3854 | 0.6693 | |
|
| 1.3853 | 0.4 | 475 | 1.3854 | 0.6615 | |
|
| 1.3876 | 0.42 | 500 | 1.3853 | 0.6753 | |
|
| 1.386 | 0.44 | 525 | 1.3853 | 0.6788 | |
|
| 1.3844 | 0.46 | 550 | 1.3852 | 0.6784 | |
|
| 1.386 | 0.48 | 575 | 1.3852 | 0.6845 | |
|
| 1.3861 | 0.5 | 600 | 1.3852 | 0.6753 | |
|
| 1.3872 | 0.52 | 625 | 1.3851 | 0.6788 | |
|
| 1.3864 | 0.55 | 650 | 1.3851 | 0.6827 | |
|
| 1.3857 | 0.57 | 675 | 1.3850 | 0.6918 | |
|
| 1.385 | 0.59 | 700 | 1.3850 | 0.6823 | |
|
| 1.3851 | 0.61 | 725 | 1.3850 | 0.6714 | |
|
| 1.3862 | 0.63 | 750 | 1.3850 | 0.6784 | |
|
| 1.3851 | 0.65 | 775 | 1.3849 | 0.6836 | |
|
| 1.3839 | 0.67 | 800 | 1.3849 | 0.6775 | |
|
| 1.3859 | 0.69 | 825 | 1.3849 | 0.6823 | |
|
| 1.3847 | 0.71 | 850 | 1.3848 | 0.6910 | |
|
| 1.3848 | 0.73 | 875 | 1.3848 | 0.6853 | |
|
| 1.3859 | 0.75 | 900 | 1.3848 | 0.6827 | |
|
| 1.3865 | 0.78 | 925 | 1.3848 | 0.6866 | |
|
| 1.3857 | 0.8 | 950 | 1.3848 | 0.6810 | |
|
| 1.3861 | 0.82 | 975 | 1.3848 | 0.6845 | |
|
| 1.3864 | 0.84 | 1000 | 1.3848 | 0.6840 | |
|
| 1.3851 | 0.86 | 1025 | 1.3848 | 0.6806 | |
|
| 1.3846 | 0.88 | 1050 | 1.3848 | 0.6823 | |
|
| 1.3866 | 0.9 | 1075 | 1.3848 | 0.6845 | |
|
| 1.3857 | 0.92 | 1100 | 1.3847 | 0.6814 | |
|
| 1.3858 | 0.94 | 1125 | 1.3847 | 0.6892 | |
|
| 1.3856 | 0.96 | 1150 | 1.3847 | 0.6810 | |
|
| 1.3853 | 0.99 | 1175 | 1.3848 | 0.6788 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.35.2 |
|
- Pytorch 2.1.0+cu118 |
|
- Datasets 2.15.0 |
|
- Tokenizers 0.15.0 |
|
|