luganda-mental-health-mixed

This model is a fine-tuned version of africanobyamugisha/luganda-asr-wav2vec2-ab1 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.6641
• Wer: 0.6617
• Cer: 0.3539

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: 4
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 300
• num_epochs: 15

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
1.9217	6.7797	50	1.7411	0.6647	0.3544
1.855	13.5593	100	1.6641	0.6617	0.3539

Framework versions

• Transformers 4.40.2
• Pytorch 2.3.1+cu121
• Datasets 2.19.2
• Tokenizers 0.19.1