m2m100_418M_it_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: 0.4792
• Bleu: 33.3519

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
3.8876	0.3676	500	0.5692	29.8373
0.5702	0.7353	1000	0.5110	31.5694
0.5284	1.1029	1500	0.4988	32.1943
0.4827	1.4706	2000	0.4896	32.6095
0.4826	1.8382	2500	0.4843	32.9797
0.4567	2.2059	3000	0.4835	33.0851
0.4461	2.5735	3500	0.4800	33.2566
0.4426	2.9412	4000	0.4792	33.3519

Framework versions

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