|
--- |
|
license: mit |
|
base_model: microsoft/git-base |
|
tags: |
|
- generated_from_trainer |
|
model-index: |
|
- name: GIT-naruto |
|
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. --> |
|
|
|
# GIT-naruto |
|
|
|
This model is a fine-tuned version of [microsoft/git-base](https://huggingface.co/microsoft/git-base) on an unknown dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.0774 |
|
- Wer Score: 16.0923 |
|
|
|
## 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 |
|
- 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 | |
|
|:-------------:|:-----:|:----:|:---------------:|:---------:| |
|
| 7.4722 | 0.93 | 50 | 4.5072 | 21.6154 | |
|
| 2.1729 | 1.85 | 100 | 0.3006 | 0.5077 | |
|
| 0.0896 | 2.78 | 150 | 0.0626 | 0.6154 | |
|
| 0.0296 | 3.7 | 200 | 0.0647 | 21.7538 | |
|
| 0.0228 | 4.63 | 250 | 0.0599 | 21.7077 | |
|
| 0.0169 | 5.56 | 300 | 0.0627 | 3.5846 | |
|
| 0.0162 | 6.48 | 350 | 0.0611 | 17.0769 | |
|
| 0.0147 | 7.41 | 400 | 0.0649 | 21.6769 | |
|
| 0.0131 | 8.33 | 450 | 0.0631 | 15.0154 | |
|
| 0.0119 | 9.26 | 500 | 0.0668 | 19.3231 | |
|
| 0.0117 | 10.19 | 550 | 0.0645 | 20.3231 | |
|
| 0.0106 | 11.11 | 600 | 0.0631 | 21.6308 | |
|
| 0.0099 | 12.04 | 650 | 0.0655 | 17.6923 | |
|
| 0.0098 | 12.96 | 700 | 0.0662 | 18.0615 | |
|
| 0.0092 | 13.89 | 750 | 0.0656 | 18.1385 | |
|
| 0.0089 | 14.81 | 800 | 0.0658 | 21.6615 | |
|
| 0.0086 | 15.74 | 850 | 0.0677 | 20.4 | |
|
| 0.0079 | 16.67 | 900 | 0.0684 | 21.6462 | |
|
| 0.0085 | 17.59 | 950 | 0.0701 | 21.6615 | |
|
| 0.0089 | 18.52 | 1000 | 0.0716 | 16.8923 | |
|
| 0.0083 | 19.44 | 1050 | 0.0685 | 21.6769 | |
|
| 0.0079 | 20.37 | 1100 | 0.0665 | 21.7077 | |
|
| 0.0075 | 21.3 | 1150 | 0.0685 | 19.5231 | |
|
| 0.0078 | 22.22 | 1200 | 0.0669 | 20.7385 | |
|
| 0.0078 | 23.15 | 1250 | 0.0677 | 18.6923 | |
|
| 0.007 | 24.07 | 1300 | 0.0698 | 19.7231 | |
|
| 0.008 | 25.0 | 1350 | 0.0682 | 20.4769 | |
|
| 0.0073 | 25.93 | 1400 | 0.0705 | 19.3231 | |
|
| 0.008 | 26.85 | 1450 | 0.0738 | 21.6615 | |
|
| 0.0071 | 27.78 | 1500 | 0.0722 | 19.9231 | |
|
| 0.0064 | 28.7 | 1550 | 0.0731 | 21.6923 | |
|
| 0.0063 | 29.63 | 1600 | 0.0741 | 20.5385 | |
|
| 0.0069 | 30.56 | 1650 | 0.0780 | 19.8462 | |
|
| 0.0063 | 31.48 | 1700 | 0.0763 | 16.9538 | |
|
| 0.0061 | 32.41 | 1750 | 0.0775 | 19.7846 | |
|
| 0.0062 | 33.33 | 1800 | 0.0772 | 19.1077 | |
|
| 0.0065 | 34.26 | 1850 | 0.0737 | 17.7231 | |
|
| 0.0062 | 35.19 | 1900 | 0.0752 | 19.5385 | |
|
| 0.0058 | 36.11 | 1950 | 0.0748 | 19.4 | |
|
| 0.006 | 37.04 | 2000 | 0.0752 | 18.4154 | |
|
| 0.0053 | 37.96 | 2050 | 0.0746 | 17.1385 | |
|
| 0.0053 | 38.89 | 2100 | 0.0766 | 15.8154 | |
|
| 0.0052 | 39.81 | 2150 | 0.0770 | 17.2 | |
|
| 0.0049 | 40.74 | 2200 | 0.0763 | 19.3538 | |
|
| 0.0051 | 41.67 | 2250 | 0.0766 | 19.9692 | |
|
| 0.0046 | 42.59 | 2300 | 0.0768 | 19.9846 | |
|
| 0.0045 | 43.52 | 2350 | 0.0773 | 16.3692 | |
|
| 0.0044 | 44.44 | 2400 | 0.0771 | 16.7846 | |
|
| 0.0041 | 45.37 | 2450 | 0.0773 | 17.6308 | |
|
| 0.0042 | 46.3 | 2500 | 0.0774 | 16.0615 | |
|
| 0.0041 | 47.22 | 2550 | 0.0767 | 16.3231 | |
|
| 0.004 | 48.15 | 2600 | 0.0771 | 16.1846 | |
|
| 0.0037 | 49.07 | 2650 | 0.0772 | 16.0462 | |
|
| 0.0035 | 50.0 | 2700 | 0.0774 | 16.0923 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.37.2 |
|
- Pytorch 2.0.1+cu117 |
|
- Datasets 2.16.1 |
|
- Tokenizers 0.15.1 |
|
|
