a56be6095abb65c0b84f9140dfadec2a

This model is a fine-tuned version of google-t5/t5-3b on the Helsinki-NLP/opus_books [es-fi] dataset. It achieves the following results on the evaluation set:

• Loss: 2.0911
• Data Size: 1.0
• Epoch Runtime: 48.1554
• Bleu: 1.1972

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	3.2546	0	3.3501	0.2423
No log	1	83	3.1034	0.0078	4.2247	0.2751
No log	2	166	2.8949	0.0156	10.0513	0.3943
No log	3	249	2.7904	0.0312	13.5162	0.5168
0.0892	4	332	2.6907	0.0625	20.3578	0.7662
0.0892	5	415	2.5995	0.125	22.6830	0.4238
0.0892	6	498	2.4772	0.25	25.1022	0.5624
0.4371	7	581	2.3501	0.5	35.2107	0.7744
2.423	8.0	664	2.2163	1.0	59.2120	0.9333
2.2971	9.0	747	2.1314	1.0	48.2086	0.9342
2.1095	10.0	830	2.0723	1.0	50.8291	0.9837
1.9711	11.0	913	2.0317	1.0	47.6761	1.0160
1.8821	12.0	996	2.0049	1.0	50.0957	0.9929
1.7285	13.0	1079	2.0002	1.0	48.6026	1.0265
1.6319	14.0	1162	1.9995	1.0	53.9101	1.0820
1.5364	15.0	1245	2.0094	1.0	54.4782	1.2083
1.4082	16.0	1328	2.0113	1.0	48.3792	1.0721
1.3177	17.0	1411	2.0655	1.0	58.8108	1.3113
1.2454	18.0	1494	2.0911	1.0	48.1554	1.1972

Framework versions

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