331e5e59db1313c5cc43edf2bff0d77d

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

• Loss: 2.1829
• Data Size: 1.0
• Epoch Runtime: 55.7004
• Bleu: 6.1623

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	6.1111	0	4.0317	1.3893
No log	1	91	5.1274	0.0078	4.6968	2.2889
No log	2	182	4.3960	0.0156	9.3351	3.6008
No log	3	273	3.9466	0.0312	14.7973	5.1294
No log	4	364	3.3690	0.0625	20.6799	7.3929
No log	5	455	3.0356	0.125	25.2738	10.1288
No log	6	546	2.6303	0.25	28.3762	12.1411
0.3946	7	637	2.2583	0.5	41.2412	4.7535
2.4956	8.0	728	2.0679	1.0	62.5970	5.4057
2.1141	9.0	819	2.0105	1.0	61.3609	5.7262
1.8301	10.0	910	2.0193	1.0	56.8296	5.8176
1.6363	11.0	1001	2.0104	1.0	57.8519	6.0920
1.4348	12.0	1092	2.0342	1.0	61.7998	6.1047
1.273	13.0	1183	2.0816	1.0	66.7489	6.1938
1.1293	14.0	1274	2.1396	1.0	55.6211	6.1618
1.0023	15.0	1365	2.1829	1.0	55.7004	6.1623

Framework versions

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