2ceaea9e502b528e9538edb1cb9746aa

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

• Loss: 3.1414
• Data Size: 1.0
• Epoch Runtime: 38.1692
• Bleu: 0.2150

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	221.5325	0	3.0682	0.0017
No log	1	74	198.9404	0.0078	4.5653	0.0030
No log	2	148	168.5888	0.0156	5.5722	0.0038
6.7009	3	222	143.3611	0.0312	7.2210	0.0039
6.7009	4	296	113.6552	0.0625	9.3702	0.0033
9.5048	5	370	64.5368	0.125	12.2335	0.0015
9.5048	6	444	28.8386	0.25	16.2880	0.0032
13.9289	7	518	16.8321	0.5	23.9353	0.0021
16.6125	8.0	592	11.4163	1.0	39.7476	0.0030
16.3322	9.0	666	9.6635	1.0	37.4544	0.0030
14.8335	10.0	740	9.1268	1.0	37.3580	0.0034
12.4568	11.0	814	8.0397	1.0	37.2919	0.0093
11.6423	12.0	888	7.3618	1.0	37.3146	0.0065
10.6014	13.0	962	7.2995	1.0	37.3232	0.0111
10.1262	14.0	1036	6.5569	1.0	37.7075	0.0083
9.2616	15.0	1110	6.1059	1.0	37.4836	0.0434
8.889	16.0	1184	5.7652	1.0	37.4712	0.0298
8.3487	17.0	1258	5.4665	1.0	37.5588	0.0884
8.0774	18.0	1332	5.2253	1.0	37.0616	0.0377
7.5902	19.0	1406	5.3379	1.0	37.2366	0.0450
7.404	20.0	1480	5.1511	1.0	37.1286	0.0331
7.0	21.0	1554	4.8344	1.0	37.0656	0.0798
6.8283	22.0	1628	4.8185	1.0	37.7522	0.0704
6.5355	23.0	1702	4.6294	1.0	37.0545	0.0906
6.4027	24.0	1776	4.4492	1.0	36.9281	0.0587
6.0659	25.0	1850	4.3216	1.0	37.2367	0.1275
5.9587	26.0	1924	4.2734	1.0	37.1520	0.1169
5.8082	27.0	1998	4.1049	1.0	37.4815	0.1064
5.6503	28.0	2072	4.0866	1.0	37.3501	0.1118
5.4901	29.0	2146	4.0389	1.0	37.5989	0.0751
5.342	30.0	2220	4.0093	1.0	37.1668	0.1192
5.2011	31.0	2294	3.8163	1.0	37.9941	0.1822
5.1137	32.0	2368	3.7344	1.0	38.1650	0.1568
5.0031	33.0	2442	3.8307	1.0	37.3691	0.0590
4.8702	34.0	2516	3.6911	1.0	37.8558	0.1390
4.8229	35.0	2590	3.7435	1.0	37.3599	0.1256
4.6841	36.0	2664	3.6212	1.0	38.0017	0.0937
4.607	37.0	2738	3.5849	1.0	37.0052	0.0807
4.4539	38.0	2812	3.6441	1.0	36.9756	0.0517
4.446	39.0	2886	3.5075	1.0	38.1396	0.0910
4.3089	40.0	2960	3.5159	1.0	36.8978	0.1166
4.2739	41.0	3034	3.3696	1.0	37.1386	0.1391
4.1601	42.0	3108	3.4174	1.0	38.0813	0.0957
4.0882	43.0	3182	3.3701	1.0	37.5271	0.1303
4.0148	44.0	3256	3.2427	1.0	37.6244	0.1631
3.9753	45.0	3330	3.2444	1.0	37.7280	0.1849
3.9001	46.0	3404	3.3593	1.0	37.2999	0.1094
3.8799	47.0	3478	3.2341	1.0	37.1080	0.1838
3.7997	48.0	3552	3.1679	1.0	37.5732	0.2302
3.7542	49.0	3626	3.1729	1.0	37.7116	0.1509
3.686	50.0	3700	3.1414	1.0	38.1692	0.2150

Framework versions

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