metadata
language:
- en
license: apache-2.0
base_model: openai/whisper-tiny.en
tags:
- generated_from_trainer
datasets:
- Dev372/Medical_STT_Dataset_1.1
metrics:
- wer
model-index:
- name: English Whisper Model
results:
- task:
name: Automatic Speech Recognition
type: automatic-speech-recognition
dataset:
name: Medical
type: Dev372/Medical_STT_Dataset_1.1
args: 'split: test'
metrics:
- name: Wer
type: wer
value: 6.283680067931677
English Whisper Model
This model is a fine-tuned version of openai/whisper-tiny.en on the Medical dataset. It achieves the following results on the evaluation set:
- Loss: 0.1252
- Wer: 6.2837
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: 18
- 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
- training_steps: 1500
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 1.2359 | 0.2825 | 100 | 1.0423 | 10.4935 |
| 0.6633 | 0.5650 | 200 | 0.6451 | 9.5072 |
| 0.4199 | 0.8475 | 300 | 0.3864 | 8.5078 |
| 0.1541 | 1.1299 | 400 | 0.1895 | 7.4202 |
| 0.1228 | 1.4124 | 500 | 0.1642 | 6.8781 |
| 0.1132 | 1.6949 | 600 | 0.1471 | 6.8422 |
| 0.1076 | 1.9774 | 700 | 0.1356 | 6.3261 |
| 0.0717 | 2.2599 | 800 | 0.1333 | 6.1334 |
| 0.0682 | 2.5424 | 900 | 0.1284 | 6.3947 |
| 0.0627 | 2.8249 | 1000 | 0.1265 | 6.5972 |
| 0.0367 | 3.1073 | 1100 | 0.1261 | 6.2478 |
| 0.0452 | 3.3898 | 1200 | 0.1265 | 6.3784 |
| 0.0374 | 3.6723 | 1300 | 0.1257 | 6.3980 |
| 0.0523 | 3.9548 | 1400 | 0.1248 | 6.1596 |
| 0.031 | 4.2373 | 1500 | 0.1252 | 6.2837 |
Framework versions
- Transformers 4.43.3
- Pytorch 2.1.2
- Datasets 2.20.0
- Tokenizers 0.19.1