d956e35f516ef9307c88f207ab37b6cd

This model is a fine-tuned version of Helsinki-NLP/opus-mt-tc-bible-big-deu_eng_fra_por_spa-mul on the Helsinki-NLP/opus_books [fi-no] dataset. It achieves the following results on the evaluation set:

• Loss: 2.6617
• Data Size: 1.0
• Epoch Runtime: 7.2854
• Bleu: 1.3124

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	8.2849	0	1.0260	0.0057
No log	1	85	6.4789	0.0078	1.8355	0.0143
No log	2	170	5.2711	0.0156	1.3322	0.0446
No log	3	255	4.4010	0.0312	1.6956	0.0753
No log	4	340	3.8542	0.0625	2.1292	0.1811
0.2833	5	425	3.4534	0.125	2.5376	0.2663
0.2833	6	510	3.1057	0.25	3.1638	0.4711
0.9721	7	595	2.8035	0.5	4.5379	0.6601
2.6315	8.0	680	2.5266	1.0	6.8046	0.8841
2.2365	9.0	765	2.4440	1.0	6.6023	1.0621
2.0002	10.0	850	2.4050	1.0	6.8575	1.1243
1.7505	11.0	935	2.4048	1.0	6.9942	1.2123
1.5317	12.0	1020	2.4563	1.0	7.3724	1.2793
1.3679	13.0	1105	2.5099	1.0	7.3836	1.2385
1.1589	14.0	1190	2.5552	1.0	8.0835	1.3212
1.0106	15.0	1275	2.6617	1.0	7.2854	1.3124

Framework versions

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