metadata
library_name: transformers
language:
- en
license: apache-2.0
base_model: openai/whisper-small.en
tags:
- generated_from_trainer
datasets:
- arlco-calls
metrics:
- wer
model-index:
- name: transcribe-arlco-calls
results:
- task:
name: Automatic Speech Recognition
type: automatic-speech-recognition
dataset:
name: arlco-calls
type: arlco-calls
args: 'config: hi, split: test'
metrics:
- name: Wer
type: wer
value: 1.9950487840396096
transcribe-arlco-calls
This model is a fine-tuned version of openai/whisper-small.en on the arlco-calls dataset. It achieves the following results on the evaluation set:
- Loss: 0.0040
- Wer: 1.9950
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: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 16
- optimizer: Use OptimizerNames.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: 20
- training_steps: 400
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.844 | 2.0 | 50 | 0.7820 | 14.4313 |
| 0.3868 | 4.0 | 100 | 0.3769 | 7.6453 |
| 0.1698 | 6.0 | 150 | 0.0941 | 5.4609 |
| 0.0241 | 8.0 | 200 | 0.0248 | 4.9949 |
| 0.0285 | 10.0 | 250 | 0.0102 | 5.1551 |
| 0.007 | 12.0 | 300 | 0.0055 | 2.2426 |
| 0.0027 | 14.0 | 350 | 0.0043 | 1.9659 |
| 0.0031 | 16.0 | 400 | 0.0040 | 1.9950 |
Framework versions
- Transformers 4.48.3
- Pytorch 2.5.1+cu124
- Datasets 3.2.0
- Tokenizers 0.21.0