wav2vec2_turkish_gender_classification

This model is a fine-tuned version of facebook/wav2vec2-base on the common_voice_17_0 dataset. It achieves the following results on the evaluation set:

• Loss: 0.7567
• Accuracy: 0.8479

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: 3e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 128
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.1723	0.9968	233	0.5006	0.8003
0.0967	1.9979	467	0.3800	0.8722
0.0696	2.9989	701	0.5256	0.8449
0.0451	4.0	935	0.5080	0.8879
0.0647	4.9968	1168	0.5977	0.8551
0.0322	5.9979	1402	0.7294	0.8463
0.0249	6.9989	1636	1.0826	0.7830
0.0189	8.0	1870	0.6995	0.8485
0.0276	8.9968	2103	0.8064	0.8360
0.0167	9.9679	2330	0.7567	0.8479

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.0+cu121
• Datasets 3.0.0
• Tokenizers 0.19.1