File size: 1,869 Bytes
00c9280 685c54a 00c9280 685c54a 00c9280 685c54a 00c9280 5e59b69 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 |
---
library_name: transformers
license: apache-2.0
base_model: openai/whisper-small
tags:
- generated_from_trainer
metrics:
- wer
model-index:
- name: w_small
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. -->
# w_small
This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.7832
- Wer: 82.6298
## 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: 1e-05
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- num_epochs: 4
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|:-------------:|:------:|:----:|:---------------:|:-------:|
| 0.9114 | 0.4548 | 1000 | 0.8773 | 80.5271 |
| 0.8239 | 0.9095 | 2000 | 0.8073 | 72.1577 |
| 0.6064 | 1.3643 | 3000 | 0.7840 | 74.4663 |
| 0.6283 | 1.8190 | 4000 | 0.7717 | 78.3562 |
| 0.5439 | 2.2738 | 5000 | 0.7827 | 78.6556 |
| 0.5574 | 2.7285 | 6000 | 0.7720 | 71.1815 |
| 0.454 | 3.1833 | 7000 | 0.7840 | 89.8216 |
| 0.4246 | 3.6380 | 8000 | 0.7832 | 82.6298 |
### Framework versions
- Transformers 4.44.2
- Pytorch 2.4.1+cu118
- Datasets 3.0.0
- Tokenizers 0.19.1
|