File size: 4,297 Bytes
1ddaaa5 |
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 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 |
---
license: apache-2.0
base_model: openai/whisper-small
tags:
- generated_from_trainer
datasets:
- common_voice_16_1
metrics:
- wer
model-index:
- name: openai/whisper-small
results:
- task:
name: Automatic Speech Recognition
type: automatic-speech-recognition
dataset:
name: common_voice_16_1
type: common_voice_16_1
config: eu
split: test
args: eu
metrics:
- name: Wer
type: wer
value: 12.912302641863374
---
<!-- 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. -->
# openai/whisper-small
This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the common_voice_16_1 dataset.
It achieves the following results on the evaluation set:
- Loss: 0.5425
- Wer: 12.9123
## 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: 64
- eval_batch_size: 32
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 256
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 40000
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|:-------------:|:------:|:-----:|:---------------:|:-------:|
| 0.0153 | 10.03 | 1000 | 0.2690 | 15.3119 |
| 0.0029 | 20.05 | 2000 | 0.3132 | 15.0334 |
| 0.0018 | 30.08 | 3000 | 0.3312 | 14.6113 |
| 0.0009 | 40.1 | 4000 | 0.3375 | 14.0916 |
| 0.0037 | 50.13 | 5000 | 0.3306 | 14.3241 |
| 0.0002 | 60.15 | 6000 | 0.3628 | 13.5464 |
| 0.0001 | 70.18 | 7000 | 0.3804 | 13.4985 |
| 0.0001 | 80.2 | 8000 | 0.3961 | 13.5298 |
| 0.0 | 90.23 | 9000 | 0.4117 | 13.5650 |
| 0.0 | 100.25 | 10000 | 0.4282 | 13.6246 |
| 0.0001 | 110.28 | 11000 | 0.3542 | 13.0061 |
| 0.0001 | 120.3 | 12000 | 0.3697 | 13.1282 |
| 0.0 | 130.33 | 13000 | 0.3874 | 12.9934 |
| 0.0 | 140.35 | 14000 | 0.4002 | 12.9582 |
| 0.0 | 150.38 | 15000 | 0.4120 | 12.9455 |
| 0.0 | 160.4 | 16000 | 0.4246 | 12.9631 |
| 0.0 | 170.43 | 17000 | 0.4369 | 13.0071 |
| 0.0 | 180.45 | 18000 | 0.4501 | 13.0364 |
| 0.0 | 190.48 | 19000 | 0.4638 | 13.0374 |
| 0.0 | 200.5 | 20000 | 0.4786 | 13.0891 |
| 0.0001 | 210.53 | 21000 | 0.3785 | 12.7374 |
| 0.0 | 220.55 | 22000 | 0.4097 | 12.8166 |
| 0.0 | 230.58 | 23000 | 0.4236 | 12.8175 |
| 0.0 | 240.6 | 24000 | 0.4340 | 12.8039 |
| 0.0 | 250.63 | 25000 | 0.4431 | 12.8156 |
| 0.0 | 260.65 | 26000 | 0.4517 | 12.8058 |
| 0.0 | 270.68 | 27000 | 0.4601 | 12.7921 |
| 0.0 | 280.7 | 28000 | 0.4689 | 12.8029 |
| 0.0 | 290.73 | 29000 | 0.4774 | 12.8039 |
| 0.0 | 300.75 | 30000 | 0.4863 | 12.7960 |
| 0.0 | 310.78 | 31000 | 0.4949 | 12.7912 |
| 0.0 | 320.8 | 32000 | 0.5037 | 12.8107 |
| 0.0 | 330.83 | 33000 | 0.5115 | 12.8087 |
| 0.0 | 340.85 | 34000 | 0.5191 | 12.8293 |
| 0.0 | 350.88 | 35000 | 0.5256 | 12.8918 |
| 0.0 | 360.9 | 36000 | 0.5313 | 12.8810 |
| 0.0 | 370.93 | 37000 | 0.5361 | 12.9045 |
| 0.0 | 380.95 | 38000 | 0.5394 | 12.8996 |
| 0.0 | 390.98 | 39000 | 0.5417 | 12.9123 |
| 0.0 | 401.0 | 40000 | 0.5425 | 12.9123 |
### Framework versions
- Transformers 4.37.2
- Pytorch 2.2.0+cu121
- Datasets 2.16.1
- Tokenizers 0.15.1
|