|
--- |
|
library_name: transformers |
|
license: apache-2.0 |
|
base_model: openai/whisper-base |
|
tags: |
|
- generated_from_trainer |
|
model-index: |
|
- name: whisper-base-zh |
|
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. --> |
|
|
|
# whisper-base-zh |
|
|
|
This model is a fine-tuned version of [openai/whisper-base](https://huggingface.co/openai/whisper-base) on the None dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.3717 |
|
- Cer: 15.8574 |
|
|
|
## 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-06 |
|
- train_batch_size: 16 |
|
- eval_batch_size: 8 |
|
- seed: 42 |
|
- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments |
|
- lr_scheduler_type: linear |
|
- lr_scheduler_warmup_steps: 100 |
|
- training_steps: 4000 |
|
- mixed_precision_training: Native AMP |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Cer | |
|
|:-------------:|:-----:|:----:|:---------------:|:-------:| |
|
| 0.5054 | 0.25 | 100 | 0.4976 | 21.1112 | |
|
| 0.4526 | 0.5 | 200 | 0.4336 | 18.1724 | |
|
| 0.4131 | 0.75 | 300 | 0.4105 | 20.9577 | |
|
| 0.3772 | 1.0 | 400 | 0.3952 | 17.7166 | |
|
| 0.297 | 1.25 | 500 | 0.3872 | 17.8054 | |
|
| 0.2837 | 1.5 | 600 | 0.3798 | 18.3740 | |
|
| 0.2801 | 1.75 | 700 | 0.3747 | 15.5887 | |
|
| 0.2776 | 2.0 | 800 | 0.3677 | 16.4739 | |
|
| 0.1981 | 2.25 | 900 | 0.3697 | 17.1169 | |
|
| 0.2198 | 2.5 | 1000 | 0.3662 | 16.7474 | |
|
| 0.2133 | 2.75 | 1100 | 0.3624 | 15.8334 | |
|
| 0.2015 | 3.0 | 1200 | 0.3597 | 15.9798 | |
|
| 0.1597 | 3.25 | 1300 | 0.3633 | 15.7902 | |
|
| 0.1796 | 3.5 | 1400 | 0.3611 | 16.7834 | |
|
| 0.145 | 3.75 | 1500 | 0.3607 | 16.6947 | |
|
| 0.1581 | 4.0 | 1600 | 0.3602 | 16.2005 | |
|
| 0.1235 | 4.25 | 1700 | 0.3639 | 14.9530 | |
|
| 0.1118 | 4.5 | 1800 | 0.3674 | 15.3344 | |
|
| 0.1266 | 4.75 | 1900 | 0.3654 | 15.3728 | |
|
| 0.1214 | 5.0 | 2000 | 0.3644 | 15.3248 | |
|
| 0.0911 | 5.25 | 2100 | 0.3678 | 15.8238 | |
|
| 0.0969 | 5.5 | 2200 | 0.3703 | 15.8046 | |
|
| 0.0956 | 5.75 | 2300 | 0.3717 | 15.8574 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.46.2 |
|
- Pytorch 2.5.1+cu121 |
|
- Datasets 3.1.0 |
|
- Tokenizers 0.20.3 |
|
|
