Ahmed107/hamsa-tiny-v0.2

Whisper Small Arabic

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

• Loss: 0.5583
• Wer: 42.7609

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: 32
• 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: 500
• training_steps: 10000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.7005	0.2	2000	0.7135	51.5366
0.6267	0.4	4000	0.6309	50.9433
0.5886	0.6	6000	0.5892	50.0225
0.5627	0.8	8000	0.5679	43.9450
0.5694	1.0	10000	0.5583	42.7609

Framework versions

• Transformers 4.38.0.dev0
• Pytorch 2.1.0+cu121
• Datasets 2.17.1.dev0
• Tokenizers 0.15.1