fd1d762135b5a3147b24497f6c0c9e19

This model is a fine-tuned version of google/long-t5-tglobal-xl on the Helsinki-NLP/opus_books [en-it] dataset. It achieves the following results on the evaluation set:

• Loss: 1.3812
• Data Size: 1.0
• Epoch Runtime: 436.5928
• Bleu: 8.9089

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	3.1391	0	32.7407	0.5970
No log	1	808	2.2314	0.0078	36.5947	2.6952
No log	2	1616	2.1153	0.0156	44.3638	3.0645
No log	3	2424	2.0180	0.0312	54.1681	4.3630
0.0794	4	3232	1.9191	0.0625	65.5805	3.9029
2.3351	5	4040	1.8165	0.125	92.0183	4.2930
2.0799	6	4848	1.7112	0.25	139.2675	4.7301
1.8298	7	5656	1.5861	0.5	241.1930	5.3961
1.6693	8.0	6464	1.4587	1.0	439.3655	6.2468
1.4766	9.0	7272	1.3798	1.0	434.6597	6.8446
1.3517	10.0	8080	1.3410	1.0	436.2038	7.5704
1.2478	11.0	8888	1.3058	1.0	435.6618	7.8762
1.1424	12.0	9696	1.2946	1.0	436.6668	8.1215
1.0611	13.0	10504	1.2936	1.0	436.7658	8.4243
0.9789	14.0	11312	1.2922	1.0	434.7293	8.6227
0.9173	15.0	12120	1.3115	1.0	433.2112	8.7182
0.8621	16.0	12928	1.3215	1.0	434.2053	8.7305
0.7782	17.0	13736	1.3393	1.0	438.3314	8.8674
0.7158	18.0	14544	1.3812	1.0	436.5928	8.9089

Framework versions

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