02f96e631041e28f81661cdc92545008

This model is a fine-tuned version of google/mt5-large on the Helsinki-NLP/opus_books [es-fr] dataset. It achieves the following results on the evaluation set:

• Loss: 1.1543
• Data Size: 1.0
• Epoch Runtime: 563.5054
• Bleu: 14.6551

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: 5e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• total_train_batch_size: 32
• total_eval_batch_size: 32
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: constant
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Data Size	Epoch Runtime	Bleu
No log	0	0	25.1392	0	43.4763	0.0162
No log	1	1407	22.3288	0.0078	47.4389	0.0117
No log	2	2814	7.0068	0.0156	53.4653	0.0579
0.3836	3	4221	4.1310	0.0312	62.7488	0.2417
3.1645	4	5628	1.8861	0.0625	80.6597	1.2357
2.1444	5	7035	1.5645	0.125	113.7415	8.2465
1.8525	6	8442	1.4423	0.25	178.3995	9.7675
1.7316	7	9849	1.3651	0.5	308.0960	11.2873
1.5217	8.0	11256	1.2787	1.0	576.8079	12.5978
1.3982	9.0	12663	1.2206	1.0	563.1243	13.0822
1.3671	10.0	14070	1.1951	1.0	575.6702	13.5539
1.26	11.0	15477	1.1747	1.0	593.7553	13.6432
1.1869	12.0	16884	1.1576	1.0	565.6605	14.0193
1.1627	13.0	18291	1.1565	1.0	563.8033	14.1323
1.1296	14.0	19698	1.1461	1.0	562.7275	14.1882
1.0725	15.0	21105	1.1423	1.0	566.5551	14.4175
1.0496	16.0	22512	1.1379	1.0	564.1802	14.5031
0.9903	17.0	23919	1.1422	1.0	570.8221	14.4280
0.9572	18.0	25326	1.1473	1.0	564.4789	14.6660
0.923	19.0	26733	1.1616	1.0	564.9559	14.5670
0.8914	20.0	28140	1.1543	1.0	563.5054	14.6551

Framework versions

• Transformers 4.57.0
• Pytorch 2.8.0+cu128
• Datasets 4.2.0
• Tokenizers 0.22.1