Kabyle_xlsr-finetuned-tamasheq.en

This model is a fine-tuned version of ad019el/Kabyle_xlsr-finetuned-tamasheq.en on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.3502
• Wer: 0.8867

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: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 200

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
6.9014	3.41	300	2.9702	1.0
2.922	6.82	600	2.7233	1.0
2.4143	10.23	900	1.3795	0.9122
1.6795	13.64	1200	1.3346	0.8810
1.5	17.05	1500	1.3412	0.9207
1.4151	20.45	1800	1.3307	0.9037
1.3298	23.86	2100	1.3467	0.8782
1.2517	27.27	2400	1.3027	0.8768
1.2162	30.68	2700	1.3134	0.8810
1.1987	34.09	3000	1.3336	0.8754
1.1706	37.5	3300	1.3502	0.8867

Framework versions

• Transformers 4.33.0
• Pytorch 2.0.1+cu118
• Datasets 2.14.4
• Tokenizers 0.13.3