69e88b75427c5b57da0bc9096c70bc6b

This model is a fine-tuned version of google/long-t5-tglobal-xl on the Helsinki-NLP/opus_books [fi-no] dataset. It achieves the following results on the evaluation set:

• Loss: 1.9816
• Data Size: 1.0
• Epoch Runtime: 62.9782
• Bleu: 1.1876

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	3.5679	0	4.3721	0.0584
No log	1	85	3.0375	0.0078	5.3873	0.0375
No log	2	170	2.7687	0.0156	9.2871	0.0840
No log	3	255	2.6347	0.0312	15.3733	0.1054
No log	4	340	2.5443	0.0625	20.3571	0.1102
0.1993	5	425	2.4398	0.125	23.7270	0.2683
0.1993	6	510	2.3298	0.25	27.7623	0.3713
0.8408	7	595	2.2112	0.5	41.8405	0.5285
2.4407	8.0	680	2.1158	1.0	63.3788	0.6352
2.2352	9.0	765	2.0510	1.0	62.6046	0.7461
2.1121	10.0	850	2.0068	1.0	58.5871	0.8303
2.0228	11.0	935	1.9786	1.0	61.3216	0.8817
1.9092	12.0	1020	1.9576	1.0	56.9923	1.0020
1.8284	13.0	1105	1.9445	1.0	60.1908	1.0769
1.7448	14.0	1190	1.9276	1.0	62.9548	1.0535
1.6732	15.0	1275	1.9263	1.0	58.6772	1.1010
1.5876	16.0	1360	1.9323	1.0	62.0639	1.1050
1.4991	17.0	1445	1.9479	1.0	59.8392	1.1192
1.4189	18.0	1530	1.9636	1.0	59.2265	1.1803
1.3422	19.0	1615	1.9816	1.0	62.9782	1.1876

Framework versions

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