metadata
license: mit
base_model: microsoft/git-base
tags:
- generated_from_trainer
model-index:
- name: git-base-instagram-cap
results: []
git-base-instagram-cap
This model is a fine-tuned version of microsoft/git-base on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.1859
- Wer Score: 1.0566
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: 32
- eval_batch_size: 32
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 64
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 50
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer Score |
|---|---|---|---|---|
| 7.3647 | 3.85 | 50 | 4.5576 | 1.0161 |
| 2.4064 | 7.69 | 100 | 0.5626 | 0.9656 |
| 0.2425 | 11.54 | 150 | 0.1644 | 0.8256 |
| 0.0894 | 15.38 | 200 | 0.1631 | 0.8623 |
| 0.0636 | 19.23 | 250 | 0.1660 | 0.8730 |
| 0.0472 | 23.08 | 300 | 0.1701 | 0.8783 |
| 0.0384 | 26.92 | 350 | 0.1743 | 0.8692 |
| 0.0327 | 30.77 | 400 | 0.1778 | 0.8814 |
| 0.0286 | 34.62 | 450 | 0.1791 | 0.8891 |
| 0.0242 | 38.46 | 500 | 0.1818 | 0.8982 |
| 0.0187 | 42.31 | 550 | 0.1831 | 0.9120 |
| 0.0141 | 46.15 | 600 | 0.1856 | 1.0092 |
| 0.012 | 50.0 | 650 | 0.1859 | 1.0566 |
Framework versions
- Transformers 4.37.0.dev0
- Pytorch 2.1.0+cu121
- Datasets 2.16.1
- Tokenizers 0.15.0