Ahmed107/whisper-tiny-ar-ft-eos

whisper-tiny-ar-ft-kws-speech-commands

This model is a fine-tuned version of arbml/whisper-tiny-ar on the Speech Commands dataset. It achieves the following results on the evaluation set:

• Loss: 1.8423
• Accuracy: 0.5204

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: 5e-05
• train_batch_size: 2
• eval_batch_size: 2
• seed: 42
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.6826	1.0	1325	0.7084	0.4966
0.7052	2.0	2650	0.6965	0.5
0.7409	3.0	3975	0.6876	0.5510
0.7077	4.0	5300	0.7214	0.5170
0.7988	5.0	6625	0.7523	0.4898
0.5818	6.0	7950	0.8118	0.5510
0.7722	7.0	9275	0.9102	0.5306
1.4165	8.0	10600	1.6832	0.5
0.7113	9.0	11925	1.6268	0.5340
0.2578	10.0	13250	1.8423	0.5204

Framework versions

• Transformers 4.48.0.dev0
• Pytorch 2.5.1+cu121
• Datasets 3.2.0
• Tokenizers 0.21.0