3ea24f2ed1b858a23541f31d599e5b34

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

• Loss: 2.2718
• Data Size: 1.0
• Epoch Runtime: 21.1311
• Bleu: 2.0758

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	4.6309	0	2.1849	0.1822
No log	1	74	4.5341	0.0078	2.2990	0.2002
No log	2	148	4.1256	0.0156	2.9168	0.2166
0.1573	3	222	3.7375	0.0312	3.9709	0.2079
0.1573	4	296	3.6161	0.0625	4.9454	0.1545
0.2639	5	370	3.4622	0.125	6.3778	0.2279
0.2639	6	444	3.3369	0.25	10.1854	0.3524
0.7481	7	518	3.1923	0.5	11.8463	0.3223
2.1397	8.0	592	3.0346	1.0	20.9981	0.5087
3.146	9.0	666	2.9258	1.0	19.1073	0.5852
3.0648	10.0	740	2.8472	1.0	20.3700	0.7541
2.9566	11.0	814	2.7775	1.0	21.0247	0.8917
2.894	12.0	888	2.7275	1.0	20.0655	0.9357
2.802	13.0	962	2.6781	1.0	20.3078	1.0024
2.7706	14.0	1036	2.6386	1.0	19.4287	1.0168
2.6821	15.0	1110	2.6058	1.0	21.8017	0.9984
2.6472	16.0	1184	2.5715	1.0	20.1687	1.0827
2.5909	17.0	1258	2.5406	1.0	20.1049	1.1494
2.5565	18.0	1332	2.5214	1.0	19.9202	1.2234
2.4964	19.0	1406	2.4951	1.0	20.0591	1.2734
2.4667	20.0	1480	2.4675	1.0	21.0128	1.2103
2.4218	21.0	1554	2.4467	1.0	21.0528	1.2792
2.3851	22.0	1628	2.4419	1.0	21.1871	1.3279
2.334	23.0	1702	2.4227	1.0	20.1782	1.3110
2.3197	24.0	1776	2.3996	1.0	20.7571	1.4217
2.2742	25.0	1850	2.3887	1.0	20.8407	1.4773
2.2534	26.0	1924	2.3755	1.0	19.8197	1.4696
2.2366	27.0	1998	2.3680	1.0	20.9092	1.5464
2.2003	28.0	2072	2.3532	1.0	19.9523	1.5650
2.1657	29.0	2146	2.3426	1.0	21.2934	1.6725
2.1374	30.0	2220	2.3273	1.0	20.8864	1.6748
2.112	31.0	2294	2.3211	1.0	21.6561	1.7425
2.0838	32.0	2368	2.3296	1.0	21.1446	1.7935
2.0718	33.0	2442	2.3094	1.0	20.0078	1.8223
2.041	34.0	2516	2.2992	1.0	19.7819	1.8350
2.0212	35.0	2590	2.3038	1.0	21.2442	1.8237
1.9988	36.0	2664	2.2849	1.0	21.4291	1.9246
1.9767	37.0	2738	2.2840	1.0	21.1425	1.9711
1.9357	38.0	2812	2.2883	1.0	20.4910	1.9415
1.9175	39.0	2886	2.2718	1.0	22.1560	1.9449
1.8986	40.0	2960	2.2752	1.0	20.7935	1.9755
1.8808	41.0	3034	2.2679	1.0	20.4600	2.0046
1.8552	42.0	3108	2.2829	1.0	20.6810	1.9760
1.8316	43.0	3182	2.2647	1.0	20.8495	2.0886
1.8148	44.0	3256	2.2536	1.0	20.6102	2.0035
1.8051	45.0	3330	2.2611	1.0	22.5918	2.0329
1.7788	46.0	3404	2.2717	1.0	22.0629	2.0431
1.7648	47.0	3478	2.2586	1.0	21.0546	2.0633
1.7331	48.0	3552	2.2718	1.0	21.1311	2.0758

Framework versions

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