541ebe02f5b1b5c46aae724cfb5a9b73

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

• Loss: 1.9800
• Data Size: 1.0
• Epoch Runtime: 54.4550
• Bleu: 5.9544

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	8.0598	0	3.4852	0.0287
No log	1	91	6.6225	0.0078	3.7232	0.0354
No log	2	182	4.3832	0.0156	8.2037	0.1431
No log	3	273	3.8458	0.0312	16.0755	0.2535
No log	4	364	3.2825	0.0625	25.2886	0.4711
No log	5	455	2.7291	0.125	23.1035	0.5379
No log	6	546	2.2008	0.25	31.0410	4.7348
0.3762	7	637	1.9935	0.5	46.5189	4.5642
2.2338	8.0	728	1.8777	1.0	59.5571	5.0543
1.9027	9.0	819	1.8375	1.0	54.3774	5.4536
1.6441	10.0	910	1.8287	1.0	58.7255	5.6488
1.4658	11.0	1001	1.8503	1.0	52.4676	5.8453
1.2786	12.0	1092	1.8656	1.0	55.1717	5.8450
1.1383	13.0	1183	1.9120	1.0	51.7027	5.9746
1.0151	14.0	1274	1.9800	1.0	54.4550	5.9544

Framework versions

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