mimba/nllb-fra2ngiemboon

This model is a fine-tuned version of mimba/nllb-fra2ngiemboon on the mimba/text2text dataset. It achieves the following results on the evaluation set:

• Loss: 0.8547
• Bleu: 23.9157
• Chrf: 45.7822
• Wer: 0.7485

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
1.0006	1.0	2405	0.8702	23.0135	44.4266	0.7693
0.9540	2.0	4810	0.8676	22.8923	44.2626	0.7676
0.9151	3.0	7215	0.8611	23.2292	44.7046	0.7619
0.8812	4.0	9620	0.8589	23.5530	45.1659	0.7639
0.8497	5.0	12025	0.8552	23.4729	45.2349	0.7576
0.8247	6.0	14430	0.8527	23.5904	45.3028	0.7569
0.7974	7.0	16835	0.8519	23.6534	45.4265	0.7499
0.7823	8.0	19240	0.8499	23.8755	45.7282	0.7578
0.7608	9.0	21645	0.8524	23.8380	45.6117	0.7479
0.7433	10.0	24050	0.8547	23.9157	45.7822	0.7485

Framework versions

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