metadata

base_model: openai/whisper-tiny
datasets:
  - fleurs
language:
  - pt
license: apache-2.0
metrics:
  - wer
tags:
  - hf-asr-leaderboard
  - generated_from_trainer
model-index:
  - name: Whisper Tiny Portuguese 5000 - Chee Li
    results:
      - task:
          type: automatic-speech-recognition
          name: Automatic Speech Recognition
        dataset:
          name: Google Fleurs
          type: fleurs
          config: pt_br
          split: None
          args: 'config: pt split: test'
        metrics:
          - type: wer
            value: 102.8207418551079
            name: Wer

Whisper Tiny Portuguese 5000 - Chee Li

This model is a fine-tuned version of openai/whisper-tiny on the Google Fleurs dataset. It achieves the following results on the evaluation set:

Loss: 0.6510
Wer: 102.8207

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: 625
training_steps: 5000
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.1445	5.0251	1000	0.5040	109.3037
0.0131	10.0503	2000	0.5788	110.2628
0.0043	15.0754	3000	0.6183	112.4207
0.0027	20.1005	4000	0.6429	109.2708
0.0022	25.1256	5000	0.6510	102.8207

Framework versions

Transformers 4.43.4
Pytorch 2.3.1+cu121
Datasets 2.20.0
Tokenizers 0.19.1