|
--- |
|
license: mit |
|
base_model: microsoft/deberta-v3-large |
|
tags: |
|
- generated_from_trainer |
|
metrics: |
|
- accuracy |
|
model-index: |
|
- name: 1_microsoft_deberta_V1.0 |
|
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. --> |
|
|
|
# 1_microsoft_deberta_V1.0 |
|
|
|
This model is a fine-tuned version of [microsoft/deberta-v3-large](https://huggingface.co./microsoft/deberta-v3-large) on the None dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.5748 |
|
- Map@3: 0.8700 |
|
- Accuracy: 0.785 |
|
|
|
## 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: 2 |
|
- eval_batch_size: 4 |
|
- seed: 42 |
|
- gradient_accumulation_steps: 21 |
|
- total_train_batch_size: 42 |
|
- 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 | Accuracy | |
|
|:-------------:|:-----:|:----:|:---------------:|:------:|:--------:| |
|
| 1.6136 | 0.02 | 25 | 1.6092 | 0.4850 | 0.335 | |
|
| 1.6116 | 0.04 | 50 | 1.6063 | 0.7250 | 0.61 | |
|
| 1.4598 | 0.05 | 75 | 1.2186 | 0.7575 | 0.63 | |
|
| 1.0137 | 0.07 | 100 | 0.9068 | 0.7908 | 0.665 | |
|
| 0.9483 | 0.09 | 125 | 0.9574 | 0.8108 | 0.69 | |
|
| 0.9619 | 0.1 | 150 | 0.8634 | 0.8183 | 0.71 | |
|
| 0.8679 | 0.12 | 175 | 0.7644 | 0.8292 | 0.73 | |
|
| 0.8594 | 0.14 | 200 | 0.8161 | 0.8067 | 0.7 | |
|
| 0.8105 | 0.16 | 225 | 0.8355 | 0.82 | 0.715 | |
|
| 0.8315 | 0.17 | 250 | 0.7381 | 0.8275 | 0.73 | |
|
| 0.8275 | 0.19 | 275 | 0.7636 | 0.8433 | 0.745 | |
|
| 0.8252 | 0.21 | 300 | 0.7196 | 0.8217 | 0.73 | |
|
| 0.7801 | 0.23 | 325 | 0.6940 | 0.8367 | 0.745 | |
|
| 0.8078 | 0.24 | 350 | 0.7185 | 0.8567 | 0.775 | |
|
| 0.7583 | 0.26 | 375 | 0.7007 | 0.8433 | 0.75 | |
|
| 0.7772 | 0.28 | 400 | 0.7032 | 0.8417 | 0.75 | |
|
| 0.8204 | 0.3 | 425 | 0.7062 | 0.8500 | 0.76 | |
|
| 0.8269 | 0.32 | 450 | 0.7082 | 0.8617 | 0.785 | |
|
| 0.7418 | 0.33 | 475 | 0.7288 | 0.8517 | 0.78 | |
|
| 0.7376 | 0.35 | 500 | 0.7021 | 0.8633 | 0.78 | |
|
| 0.7519 | 0.37 | 525 | 0.6943 | 0.8642 | 0.785 | |
|
| 0.7469 | 0.39 | 550 | 0.6807 | 0.8725 | 0.805 | |
|
| 0.7244 | 0.4 | 575 | 0.6622 | 0.8692 | 0.79 | |
|
| 0.7297 | 0.42 | 600 | 0.6783 | 0.8583 | 0.775 | |
|
| 0.7259 | 0.44 | 625 | 0.6788 | 0.8550 | 0.765 | |
|
| 0.6893 | 0.46 | 650 | 0.6571 | 0.8625 | 0.785 | |
|
| 0.6871 | 0.47 | 675 | 0.6587 | 0.8492 | 0.76 | |
|
| 0.7003 | 0.49 | 700 | 0.6485 | 0.8683 | 0.785 | |
|
| 0.7094 | 0.51 | 725 | 0.6320 | 0.8675 | 0.795 | |
|
| 0.7052 | 0.53 | 750 | 0.6554 | 0.8583 | 0.78 | |
|
| 0.6873 | 0.54 | 775 | 0.6121 | 0.8550 | 0.775 | |
|
| 0.6152 | 0.56 | 800 | 0.6060 | 0.8675 | 0.785 | |
|
| 0.6741 | 0.58 | 825 | 0.6191 | 0.8808 | 0.815 | |
|
| 0.7098 | 0.59 | 850 | 0.6213 | 0.8817 | 0.815 | |
|
| 0.7029 | 0.61 | 875 | 0.6533 | 0.8725 | 0.79 | |
|
| 0.6489 | 0.63 | 900 | 0.6127 | 0.8667 | 0.79 | |
|
| 0.7289 | 0.65 | 925 | 0.6261 | 0.8750 | 0.81 | |
|
| 0.6589 | 0.67 | 950 | 0.6019 | 0.8708 | 0.8 | |
|
| 0.6876 | 0.68 | 975 | 0.6076 | 0.8725 | 0.805 | |
|
| 0.6624 | 0.7 | 1000 | 0.5810 | 0.8708 | 0.79 | |
|
| 0.6746 | 0.72 | 1025 | 0.5922 | 0.8708 | 0.79 | |
|
| 0.6644 | 0.73 | 1050 | 0.5827 | 0.8675 | 0.785 | |
|
| 0.668 | 0.75 | 1075 | 0.5814 | 0.8725 | 0.795 | |
|
| 0.6115 | 0.77 | 1100 | 0.5680 | 0.8750 | 0.8 | |
|
| 0.6799 | 0.79 | 1125 | 0.5767 | 0.8750 | 0.8 | |
|
| 0.6466 | 0.81 | 1150 | 0.5700 | 0.8725 | 0.795 | |
|
| 0.6765 | 0.82 | 1175 | 0.5700 | 0.8717 | 0.79 | |
|
| 0.6936 | 0.84 | 1200 | 0.5758 | 0.8683 | 0.785 | |
|
| 0.6239 | 0.86 | 1225 | 0.5748 | 0.8700 | 0.785 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.32.1 |
|
- Pytorch 2.0.0 |
|
- Datasets 2.9.0 |
|
- Tokenizers 0.13.3 |
|
|
