metadata
license: apache-2.0
base_model: openai/whisper-tiny
tags:
- ASR
- Papiamentu
- Whisper
- Speech Recognition
- generated_from_trainer
datasets:
- sonnygeorge/papi_asr_corpus
metrics:
- wer
model-index:
- name: Whisper Tiny Papiamentu
results:
- task:
name: Automatic Speech Recognition
type: automatic-speech-recognition
dataset:
name: Papi ASR
type: sonnygeorge/papi_asr_corpus
metrics:
- name: Wer
type: wer
value: 30.326720259606166
Whisper Tiny Papiamentu
This model is a fine-tuned version of openai/whisper-tiny on the Papi ASR dataset. It achieves the following results on the evaluation set:
- Loss: 0.3275
- Wer: 30.3267
- Cer: 15.4823
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: 3
- training_steps: 200
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
|---|---|---|---|---|---|
| 0.1932 | 0.07 | 50 | 0.3766 | 42.1934 | 28.1746 |
| 0.2235 | 0.15 | 100 | 0.3513 | 33.0998 | 17.9645 |
| 0.2989 | 0.22 | 150 | 0.3337 | 30.2898 | 14.8959 |
| 0.3269 | 0.3 | 200 | 0.3275 | 30.3267 | 15.4823 |
Framework versions
- Transformers 4.37.0.dev0
- Pytorch 2.1.0+cu121
- Datasets 2.15.0
- Tokenizers 0.15.0