metadata

language:
  - es
license: apache-2.0
base_model: openai/whisper-tiny
tags:
  - hf-asr-leaderboard
  - generated_from_trainer
datasets:
  - ctam8736/papi_asr
metrics:
  - wer
model-index:
  - name: Whisper Small Papi/Es
    results:
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: Papi ASR Test
          type: ctam8736/papi_asr
          args: 'config: es, split: test'
        metrics:
          - name: Wer
            type: wer
            value: 22.26516425326875

Whisper Small Papi/Es

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

Loss: 0.2038
Wer: 22.2652

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: 500
training_steps: 1000
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.8909	0.34	200	0.8337	59.4875
0.4264	0.69	400	0.4114	33.3977
0.2452	1.03	600	0.2663	26.1676
0.1953	1.38	800	0.2188	24.0369
0.1699	1.72	1000	0.2038	22.2652

Framework versions

Transformers 4.37.0.dev0
Pytorch 2.0.1+cu117
Datasets 2.15.0
Tokenizers 0.15.0