djoe2024sky/whisper-medium-ha-test

Whisper Small ha Test- Dinaka Ezeani

This model is a fine-tuned version of openai/whisper-medium on the Common Voice 11.0 dataset. It achieves the following results on the evaluation set:

• Loss: 1.4394

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: 0.001
• train_batch_size: 64
• eval_batch_size: 32
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 2
• training_steps: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
3.5742	0.0323	1	2.8898
3.6442	0.0645	2	2.5929
3.3101	0.0968	3	2.1882
2.797	0.1290	4	1.9838
2.6471	0.1613	5	1.7828
2.3832	0.1935	6	1.6560
2.1674	0.2258	7	1.5718
2.0717	0.2581	8	1.5056
2.0157	0.2903	9	1.4620
1.9508	0.3226	10	1.4394

Framework versions

• PEFT 0.12.0
• Transformers 4.44.2
• Pytorch 2.4.1+cu121
• Datasets 3.0.0
• Tokenizers 0.19.1