README.md

metadata

language:
  - th
license: apache-2.0
base_model: openai/whisper-tiny
tags:
  - generated_from_trainer
datasets:
  - Porameht/processed-smarthome-th
metrics:
  - wer
model-index:
  - name: whisper-tiny-smarthome-th
    results:
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: processed-smarthome-th
          type: Porameht/processed-smarthome-th
          args: 'config: th, split: test'
        metrics:
          - name: Wer
            type: wer
            value: 24.375

whisper-tiny-smarthome-th

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

• Loss: 0.0546
• Wer: 24.375

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: 32
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 5
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0072	4.1667	500	0.0546	24.375

Framework versions

• Transformers 4.41.1
• Pytorch 2.3.0+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1