README.md

metadata

language:
  - dv
license: apache-2.0
base_model: openai/whisper-small
tags:
  - generated_from_trainer
datasets:
  - mozilla-foundation/common_voice_13_0
metrics:
  - wer
model-index:
  - name: Whisper Small Dv - Ruhullah Shaikh
    results:
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: Common Voice 13
          type: mozilla-foundation/common_voice_13_0
          config: dv
          split: test
          args: dv
        metrics:
          - name: Wer
            type: wer
            value: 10.97645790590117

Whisper Small Dv - Ruhullah Shaikh

This model is a fine-tuned version of openai/whisper-small on the Common Voice 13 dataset. It achieves the following results on the evaluation set:

• Loss: 0.3049
• Wer Ortho: 57.3995
• Wer: 10.9765

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
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 50
• training_steps: 4000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer Ortho	Wer
0.1224	1.6313	500	0.1725	63.0197	13.4872
0.0448	3.2626	1000	0.1690	58.1378	11.5189
0.0297	4.8940	1500	0.1814	60.0251	11.5450
0.006	6.5253	2000	0.2352	58.2701	11.3503
0.0018	8.1566	2500	0.2639	58.3676	11.1364
0.0008	9.7879	3000	0.2888	57.7686	11.0738
0.0002	11.4192	3500	0.3015	57.3369	10.9938
0.0002	13.0506	4000	0.3049	57.3995	10.9765

Framework versions

• Transformers 4.41.2
• Pytorch 2.4.0+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1