elmamounedieye/wolof-finetuned

nllb-200-1.3B-wol-fr

This model is a fine-tuned version of nllb-200-1.3B on the elmamounedieye/agri_wol dataset. It achieves the following results on the evaluation set:

• Loss: 0.2740
• Bleu: 24.9828

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: 0.0002
• train_batch_size: 8
• eval_batch_size: 4
• seed: 42
• optimizer: Use OptimizerNames.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_steps: 100
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu
0.1963	1.0	1125	0.1795	20.8754
0.1055	2.0	2250	0.1807	21.3156
0.0422	3.0	3375	0.2031	22.9941
0.0216	4.0	4500	0.2324	22.2155
0.012	5.0	5625	0.2412	23.8844
0.0069	6.0	6750	0.2501	23.5372
0.0043	7.0	7875	0.2587	23.4568
0.0024	8.0	9000	0.2657	24.7322
0.001	9.0	10125	0.2683	24.9165
0.0006	10.0	11250	0.2740	24.9828

Framework versions

• PEFT 0.14.0
• Transformers 4.48.3
• Pytorch 2.6.0+cu124
• Datasets 3.4.1
• Tokenizers 0.21.1