f520900cd9c1cd041087e12c29c1e445

This model is a fine-tuned version of google/umt5-xl on the Helsinki-NLP/opus_books [fi-no] dataset. It achieves the following results on the evaluation set:

• Loss: 2.2454
• Data Size: 1.0
• Epoch Runtime: 59.4839
• Bleu: 8.2202

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	6.0653	0	3.8028	1.9468
No log	1	85	5.0193	0.0078	4.5767	4.0222
No log	2	170	4.4826	0.0156	11.8299	5.9615
No log	3	255	4.2189	0.0312	17.7330	6.9025
No log	4	340	3.6464	0.0625	24.1929	8.3983
0.3276	5	425	3.2718	0.125	23.1157	11.7699
0.3276	6	510	2.7775	0.25	31.1374	13.5054
1.063	7	595	2.3824	0.5	45.3406	6.6626
2.6936	8.0	680	2.1728	1.0	59.9517	7.6726
2.2218	9.0	765	2.1041	1.0	53.9321	8.0869
1.9193	10.0	850	2.0846	1.0	57.8026	8.2728
1.733	11.0	935	2.1063	1.0	53.2315	8.2971
1.5248	12.0	1020	2.1427	1.0	56.9248	8.2985
1.3702	13.0	1105	2.2019	1.0	54.9972	8.3117
1.1993	14.0	1190	2.2454	1.0	59.4839	8.2202

Framework versions

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