b15b6753225b2dd26cb377a67098d3b8

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

• Loss: 6.5205
• Data Size: 1.0
• Epoch Runtime: 19.4832
• Bleu: 0.3750

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	237.5808	0	1.7565	0.0047
No log	1	29	219.4165	0.0078	2.4461	0.0186
No log	2	58	206.7268	0.0156	3.7988	0.0138
No log	3	87	195.5772	0.0312	5.1655	0.0133
No log	4	116	175.3125	0.0625	6.7078	0.0181
No log	5	145	143.0906	0.125	9.3192	0.0199
17.0186	6	174	94.6019	0.25	11.9287	0.0097
17.0186	7	203	42.3782	0.5	14.1822	0.0037
17.0186	8.0	232	20.2820	1.0	21.5922	0.0067
35.4455	9.0	261	15.4973	1.0	19.8251	0.0189
35.4455	10.0	290	13.6906	1.0	20.5325	0.0165
22.7957	11.0	319	12.3677	1.0	20.1372	0.0335
22.7957	12.0	348	10.9040	1.0	19.5501	0.0352
18.4549	13.0	377	11.0217	1.0	20.4877	0.0181
15.8757	14.0	406	9.5505	1.0	19.9176	0.0425
15.8757	15.0	435	9.7615	1.0	20.4831	0.0280
14.4684	16.0	464	9.1320	1.0	19.5799	0.0413
14.4684	17.0	493	8.4326	1.0	19.8343	0.0903
13.1074	18.0	522	8.3413	1.0	20.3576	0.1056
12.2344	19.0	551	8.3180	1.0	19.6270	0.0970
12.2344	20.0	580	7.9973	1.0	20.2316	0.0740
11.3743	21.0	609	7.5731	1.0	19.8370	0.2492
11.3743	22.0	638	7.3466	1.0	19.5481	0.2408
10.6951	23.0	667	7.3575	1.0	19.7752	0.1385
10.6951	24.0	696	6.5999	1.0	19.4145	0.3983
10.0869	25.0	725	6.3966	1.0	19.8224	0.3704
9.5201	26.0	754	6.4392	1.0	19.5388	0.3759
9.5201	27.0	783	6.4110	1.0	20.3876	0.3772
8.9994	28.0	812	6.5906	1.0	19.3776	0.3860
8.9994	29.0	841	6.5205	1.0	19.4832	0.3750

Framework versions

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