metadata
base_model: Samuael/geez_t5-15k
tags:
- generated_from_trainer
metrics:
- wer
- bleu
model-index:
- name: geez_t5-15k
results: []
geez_t5-15k
This model is a fine-tuned version of Samuael/geez_t5-15k on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.4822
- Wer: 0.2284
- Cer: 0.0780
- Bleu: 67.0096
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: 0.0002
- train_batch_size: 64
- eval_batch_size: 128
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 10
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer | Cer | Bleu |
|---|---|---|---|---|---|---|
| 0.2496 | 1.0 | 313 | 0.3291 | 0.2456 | 0.1077 | 67.2734 |
| 0.2573 | 2.0 | 626 | 0.3445 | 0.2488 | 0.1114 | 67.1612 |
| 0.2368 | 3.0 | 939 | 0.3483 | 0.2075 | 0.0698 | 69.4248 |
| 0.1459 | 4.0 | 1252 | 0.3815 | 0.2212 | 0.0806 | 68.5281 |
| 0.1552 | 5.0 | 1565 | 0.4081 | 0.2372 | 0.0934 | 66.8604 |
| 0.1031 | 6.0 | 1878 | 0.4143 | 0.2084 | 0.0692 | 69.6668 |
| 0.1067 | 7.0 | 2191 | 0.4325 | 0.2305 | 0.0853 | 67.3479 |
| 0.1111 | 8.0 | 2504 | 0.4380 | 0.2070 | 0.0638 | 68.9579 |
| 0.0965 | 9.0 | 2817 | 0.4541 | 0.2195 | 0.0735 | 67.8519 |
| 0.0843 | 10.0 | 3130 | 0.4822 | 0.2284 | 0.0780 | 67.0096 |
Framework versions
- Transformers 4.38.2
- Pytorch 2.2.1+cu121
- Datasets 2.18.0
- Tokenizers 0.15.2