64264194fc2a4f42ecc1f95eec56730e

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

• Loss: 0.9175
• Data Size: 1.0
• Epoch Runtime: 39.3368
• Bleu: 18.8538

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	6.7874	0	1.8337	0.0492
No log	1	35	5.7017	0.0078	2.5081	0.0714
No log	2	70	4.2507	0.0156	6.4993	0.1528
No log	3	105	3.2339	0.0312	14.5468	0.4578
No log	4	140	2.5033	0.0625	22.9918	1.0598
No log	5	175	2.0593	0.125	28.6541	1.6707
No log	6	210	1.6270	0.25	22.2383	2.1778
No log	7	245	1.1479	0.5	25.1250	3.0213
0.5296	8.0	280	0.9210	1.0	41.2080	15.8497
1.2825	9.0	315	0.8598	1.0	42.7891	16.0894
0.9729	10.0	350	0.8443	1.0	30.9760	16.9181
0.9729	11.0	385	0.8413	1.0	33.6040	16.8035
0.7696	12.0	420	0.8547	1.0	38.2172	17.3521
0.6304	13.0	455	0.8768	1.0	31.0604	18.1505
0.6304	14.0	490	0.9103	1.0	33.5337	17.0979
0.5147	15.0	525	0.9175	1.0	39.3368	18.8538

Framework versions

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