1dd0abdb86a2b42c2dea0b7ec32874de

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

• Loss: 2.1759
• Data Size: 1.0
• Epoch Runtime: 41.5750
• Bleu: 1.0114

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	3.5396	0	3.2654	0.2684
No log	1	83	3.4305	0.0078	3.9074	0.2428
No log	2	166	3.2767	0.0156	5.4522	0.2444
No log	3	249	3.1750	0.0312	7.3024	0.3524
0.1016	4	332	3.0698	0.0625	9.2434	0.5500
0.1016	5	415	2.9732	0.125	12.7490	0.4256
0.1016	6	498	2.8521	0.25	18.3018	0.3553
0.504	7	581	2.7120	0.5	24.7525	0.4520
2.851	8.0	664	2.5704	1.0	44.4937	0.7189
2.7501	9.0	747	2.4680	1.0	43.2597	0.7969
2.5996	10.0	830	2.4121	1.0	41.8650	0.8310
2.4899	11.0	913	2.3604	1.0	43.0643	0.8401
2.4422	12.0	996	2.3238	1.0	43.4102	0.8580
2.3467	13.0	1079	2.2864	1.0	40.6000	0.7403
2.288	14.0	1162	2.2628	1.0	42.2209	0.8138
2.2386	15.0	1245	2.2310	1.0	41.1208	0.8524
2.1624	16.0	1328	2.2156	1.0	41.4326	0.9501
2.1103	17.0	1411	2.2142	1.0	42.8475	0.9795
2.0791	18.0	1494	2.1907	1.0	40.5784	0.9265
2.0065	19.0	1577	2.1896	1.0	41.1820	0.9252
1.9697	20.0	1660	2.1798	1.0	42.0205	0.9593
1.914	21.0	1743	2.1695	1.0	41.1141	0.9670
1.8667	22.0	1826	2.1631	1.0	42.0923	1.0171
1.8287	23.0	1909	2.1648	1.0	41.5415	0.9918
1.7956	24.0	1992	2.1665	1.0	40.2423	0.9394
1.7408	25.0	2075	2.1748	1.0	40.3875	0.9948
1.7082	26.0	2158	2.1759	1.0	41.5750	1.0114

Framework versions

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