|
--- |
|
license: apache-2.0 |
|
base_model: facebook/bart-base |
|
tags: |
|
- generated_from_trainer |
|
metrics: |
|
- wer |
|
model-index: |
|
- name: bart-kurd-spell-base |
|
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. --> |
|
|
|
# bart-kurd-spell-base |
|
|
|
This model is a fine-tuned version of [facebook/bart-base](https://huggingface.co/facebook/bart-base) on the None dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.1930 |
|
- Cer: 1.5424 |
|
- Wer: 8.3088 |
|
- Gen Len: 12.6945 |
|
|
|
## 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: 256 |
|
- eval_batch_size: 256 |
|
- seed: 42 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: linear |
|
- num_epochs: 10.0 |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Cer | Wer | Gen Len | |
|
|:-------------:|:-----:|:------:|:---------------:|:------:|:-------:|:-------:| |
|
| 0.4548 | 1.54 | 20000 | 0.4117 | 2.8856 | 13.6181 | 12.7807 | |
|
| 0.2723 | 3.07 | 40000 | 0.2736 | 2.1004 | 10.5883 | 12.6808 | |
|
| 0.2246 | 4.61 | 60000 | 0.2303 | 1.8035 | 9.4897 | 12.7048 | |
|
| 0.1812 | 6.14 | 80000 | 0.2122 | 1.6804 | 8.9349 | 12.6937 | |
|
| 0.1693 | 7.68 | 100000 | 0.2001 | 1.589 | 8.5464 | 12.7045 | |
|
| 0.1498 | 9.22 | 120000 | 0.1942 | 1.5546 | 8.3598 | 12.6935 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.34.0.dev0 |
|
- Pytorch 1.13.1 |
|
- Datasets 2.14.5 |
|
- Tokenizers 0.13.3 |
|
|
