metadata
license: apache-2.0
base_model: openai/whisper-large-v2
tags:
- generated_from_trainer
datasets:
- wanasash/enwaucymraeg
metrics:
- wer
model-index:
- name: whisper-large-v2-ec
results:
- task:
name: Automatic Speech Recognition
type: automatic-speech-recognition
dataset:
name: wanasash/enwaucymraeg default
type: wanasash/enwaucymraeg
args: default
metrics:
- name: Wer
type: wer
value: 0.21671018276762402
language:
- cy
whisper-large-v2-ec
This model is a fine-tuned version of openai/whisper-large-v2 on the wanasash/enwaucymraeg default dataset. It achieves the following results on the evaluation set:
- Loss: 0.5119
- Wer: 0.2167
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: 16
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 5000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.0112 | 13.6054 | 1000 | 0.3912 | 0.2395 |
| 0.0004 | 27.2109 | 2000 | 0.4532 | 0.2245 |
| 0.0002 | 40.8163 | 3000 | 0.4882 | 0.2175 |
| 0.0001 | 54.4218 | 4000 | 0.5051 | 0.2148 |
| 0.0001 | 68.0272 | 5000 | 0.5119 | 0.2167 |
Framework versions
- Transformers 4.44.0
- Pytorch 2.4.0+cu121
- Datasets 2.20.0
- Tokenizers 0.19.1