metadata
license: apache-2.0
tags:
- whisper-event
- generated_from_trainer
datasets:
- google/fleurs
metrics:
- wer
base_model: openai/whisper-small
model-index:
- name: Whisper Small Chinese Base
results:
- task:
type: automatic-speech-recognition
name: Automatic Speech Recognition
dataset:
name: google/fleurs cmn_hans_cn
type: google/fleurs
config: cmn_hans_cn
split: test
args: cmn_hans_cn
metrics:
- type: wer
value: 16.643891773708663
name: Wer
Whisper Small Chinese Base
This model is a fine-tuned version of openai/whisper-small on the google/fleurs cmn_hans_cn dataset. It achieves the following results on the evaluation set:
- Loss: 0.3573
- Wer: 16.6439
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
- 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 |
---|---|---|---|---|
0.0005 | 76.0 | 1000 | 0.3573 | 16.6439 |
0.0002 | 153.0 | 2000 | 0.3897 | 16.9749 |
0.0001 | 230.0 | 3000 | 0.4125 | 17.2330 |
0.0001 | 307.0 | 4000 | 0.4256 | 17.2451 |
0.0001 | 384.0 | 5000 | 0.4330 | 17.2300 |
Framework versions
- Transformers 4.26.0.dev0
- Pytorch 1.13.1+cu117
- Datasets 2.7.1.dev0
- Tokenizers 0.13.2