mimba/nllb-ngiemboon2fra

This model is a fine-tuned version of mimba/nllb-ngiemboon2fra on the mimba/text2text dataset. It achieves the following results on the evaluation set:

• Loss: 0.7438
• Bleu: 36.0431
• Chrf: 54.1025
• Wer: 0.6336

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: 2e-05
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED 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: 961
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Chrf	Wer
0.8913	1.0	2405	0.7217	37.4670	54.9183	0.6222
0.8383	2.0	4810	0.7325	37.0391	54.7729	0.6331
0.7954	3.0	7215	0.7438	36.0431	54.1025	0.6336

Framework versions

• Transformers 5.0.0
• Pytorch 2.10.0+cu128
• Datasets 4.5.0
• Tokenizers 0.22.2