M2M100-finetuned-ES-ZAP-v1

This model is a fine-tuned version of facebook/m2m100_418M on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.3271
• Bleu: 0.3862
• Meteor: 0.1099
• Ter: 165.8363

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: 1e-05
• train_batch_size: 8
• 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
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Meteor	Ter
7.1271	1.0	102	5.5135	0.0355	0.0440	888.6121
4.7892	2.0	204	3.8860	0.2755	0.0852	193.0605
3.2415	3.0	306	2.5041	0.1740	0.0855	309.2527
1.9994	4.0	408	1.4749	0.6447	0.1013	111.2100
1.1483	5.0	510	0.8498	0.3108	0.0987	183.0961
0.6665	6.0	612	0.5500	0.4980	0.0960	150.7117
0.4394	7.0	714	0.4174	0.5711	0.1068	164.4128
0.3368	8.0	816	0.3597	0.5310	0.1088	168.6833
0.2821	9.0	918	0.3348	0.5948	0.1114	161.7438
0.2644	10.0	1020	0.3271	0.3862	0.1099	165.8363

Framework versions

• Transformers 4.49.0
• Pytorch 2.6.0+cu124
• Datasets 3.4.1
• Tokenizers 0.21.1