m2m100_418M_fr_formal

This model is a fine-tuned version of facebook/m2m100_418M on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.3635
• Bleu: 41.9288

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: 32
• eval_batch_size: 8
• 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: 500
• num_epochs: 3.0

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu
3.8632	0.3554	500	0.4614	34.2520
0.4495	0.7107	1000	0.3976	38.3127
0.4137	1.0661	1500	0.3816	39.7176
0.373	1.4215	2000	0.3763	40.2181
0.3693	1.7768	2500	0.3693	41.0068
0.357	2.1322	3000	0.3675	41.4021
0.335	2.4876	3500	0.3645	41.6258
0.3336	2.8429	4000	0.3635	41.9288

Framework versions

• Transformers 4.48.3
• Pytorch 2.5.1+cu124
• Datasets 3.3.2
• Tokenizers 0.21.0