a55f5de94c289243ea7f5b22be930596

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

• Loss: 2.0482
• Data Size: 1.0
• Epoch Runtime: 38.5429
• Bleu: 2.8624

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	4.1119	0	3.1193	0.2144
No log	1	74	3.9589	0.0078	3.7150	0.2346
No log	2	148	3.3321	0.0156	5.2219	0.3324
0.1352	3	222	3.1186	0.0312	7.9675	0.4616
0.1352	4	296	2.9632	0.0625	9.8806	0.7536
0.2201	5	370	2.8418	0.125	12.8331	0.9133
0.2201	6	444	2.7179	0.25	16.7080	1.0366
0.6174	7	518	2.5883	0.5	22.9964	0.7716
1.7466	8.0	592	2.4355	1.0	38.6028	1.1599
2.4978	9.0	666	2.3423	1.0	37.9237	1.3622
2.4076	10.0	740	2.2758	1.0	38.4024	1.5715
2.2888	11.0	814	2.2214	1.0	38.4786	1.7239
2.2112	12.0	888	2.1839	1.0	38.3051	1.8597
2.101	13.0	962	2.1455	1.0	36.1867	1.8599
2.0638	14.0	1036	2.1136	1.0	37.0464	2.0712
1.9577	15.0	1110	2.0953	1.0	39.8808	2.1943
1.9117	16.0	1184	2.0838	1.0	36.4446	2.1814
1.8403	17.0	1258	2.0642	1.0	39.0805	2.2443
1.7974	18.0	1332	2.0671	1.0	38.9135	2.3334
1.7339	19.0	1406	2.0612	1.0	37.1128	2.4690
1.6825	20.0	1480	2.0455	1.0	41.0984	2.4958
1.6272	21.0	1554	2.0384	1.0	38.0693	2.5956
1.5797	22.0	1628	2.0517	1.0	38.6636	2.6271
1.5208	23.0	1702	2.0343	1.0	37.4619	2.6566
1.4857	24.0	1776	2.0297	1.0	37.8110	2.6490
1.4415	25.0	1850	2.0412	1.0	37.4855	2.7629
1.4072	26.0	1924	2.0330	1.0	39.4528	2.7936
1.3842	27.0	1998	2.0452	1.0	39.0299	2.8494
1.3323	28.0	2072	2.0482	1.0	38.5429	2.8624

Framework versions

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