702802ca759e8c681985de9fada58d4e

This model is a fine-tuned version of google/umt5-xl on the Helsinki-NLP/opus_books [fi-pl] dataset. It achieves the following results on the evaluation set:

• Loss: 2.5215
• Data Size: 1.0
• Epoch Runtime: 51.9528
• Bleu: 3.5254

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	6.2096	0	3.2886	0.6897
No log	1	70	5.5113	0.0078	3.7767	1.9698
No log	2	140	4.9102	0.0156	8.9514	2.7879
No log	3	210	4.5506	0.0312	14.1241	3.2300
No log	4	280	4.1249	0.0625	20.2292	3.6240
No log	5	350	3.7456	0.125	23.0716	3.9856
No log	6	420	3.3007	0.25	28.2285	5.4771
0.6568	7	490	2.8853	0.5	40.3148	2.8012
3.1097	8.0	560	2.5779	1.0	51.7743	2.9091
2.7649	9.0	630	2.4674	1.0	49.0130	3.1892
2.4045	10.0	700	2.4262	1.0	51.6519	3.4296
2.1527	11.0	770	2.4359	1.0	48.5027	3.6072
2.0183	12.0	840	2.4477	1.0	47.3349	3.5618
1.764	13.0	910	2.4811	1.0	49.4465	3.5486
1.5986	14.0	980	2.5215	1.0	51.9528	3.5254

Framework versions

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