README.md

metadata

license: apache-2.0
base_model: openai/whisper-tiny
tags:
  - generated_from_trainer
datasets:
  - Ransaka/SinhalaASR
metrics:
  - wer
model-index:
  - name: whisper-tiny-sinhala-20k
    results:
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: sinhala_asr
          type: sinhala_asr
          config: default
          split: test
          args: default
        metrics:
          - name: Wer
            type: wer
            value: 92.99603723159156

whisper-tiny-sinhala-20k

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

• Loss: 0.2433
• Wer: 92.9960

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: 16
• 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
0.4207	0.4	1000	0.3978	221.9058
0.2966	0.8	2000	0.3009	136.3423
0.226	1.2	3000	0.2661	97.6638
0.2224	1.6	4000	0.2510	92.3279
0.2034	2.0	5000	0.2433	92.9960

Framework versions

• Transformers 4.36.0.dev0
• Pytorch 2.0.0
• Datasets 2.1.0
• Tokenizers 0.15.0