|
--- |
|
license: mit |
|
tags: |
|
- generated_from_trainer |
|
model-index: |
|
- name: sinhala-book-captioning-repo |
|
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. --> |
|
|
|
# sinhala-book-captioning-repo |
|
|
|
This model is a fine-tuned version of [microsoft/git-base](https://huggingface.co/microsoft/git-base) on the None dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.0053 |
|
- Wer Score: 1.3218 |
|
|
|
## 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: 5e-05 |
|
- train_batch_size: 1 |
|
- eval_batch_size: 1 |
|
- seed: 42 |
|
- gradient_accumulation_steps: 2 |
|
- total_train_batch_size: 2 |
|
- 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 | Wer Score | |
|
|:-------------:|:-----:|:----:|:---------------:|:---------:| |
|
| 5.2519 | 0.3 | 50 | 4.5638 | 1.0 | |
|
| 0.6662 | 0.6 | 100 | 0.3785 | 1.0 | |
|
| 0.0214 | 0.9 | 150 | 0.0154 | 0.8161 | |
|
| 0.0123 | 1.2 | 200 | 0.0104 | 0.7356 | |
|
| 0.0077 | 1.5 | 250 | 0.0095 | 0.7126 | |
|
| 0.0091 | 1.8 | 300 | 0.0080 | 0.6897 | |
|
| 0.009 | 2.1 | 350 | 0.0076 | 0.6897 | |
|
| 0.008 | 2.4 | 400 | 0.0076 | 0.7356 | |
|
| 0.0066 | 2.69 | 450 | 0.0071 | 0.6667 | |
|
| 0.0065 | 2.99 | 500 | 0.0068 | 0.6437 | |
|
| 0.0068 | 3.29 | 550 | 0.0063 | 0.6207 | |
|
| 0.0066 | 3.59 | 600 | 0.0063 | 5.1494 | |
|
| 0.0063 | 3.89 | 650 | 0.0061 | 0.6322 | |
|
| 0.0063 | 4.19 | 700 | 0.0055 | 0.5287 | |
|
| 0.007 | 4.49 | 750 | 0.0053 | 0.5172 | |
|
| 0.0054 | 4.79 | 800 | 0.0053 | 1.3218 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.30.2 |
|
- Pytorch 1.13.0 |
|
- Datasets 2.11.0 |
|
- Tokenizers 0.13.0 |
|
|
