a3114f901017637fa4ac99911bd37b4f

This model is a fine-tuned version of google/mt5-large on the Helsinki-NLP/opus_books [es-ru] dataset. It achieves the following results on the evaluation set:

• Loss: 1.7294
• Data Size: 1.0
• Epoch Runtime: 183.5361
• Bleu: 7.5246

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	24.2166	0	13.0999	0.0155
No log	1	419	24.0466	0.0078	14.6959	0.0115
No log	2	838	21.8030	0.0156	17.2619	0.0118
0.6892	3	1257	16.8624	0.0312	21.9391	0.0098
0.6892	4	1676	13.1123	0.0625	27.9811	0.0078
1.1343	5	2095	8.3941	0.125	39.2108	0.0111
0.5339	6	2514	2.3682	0.25	58.5343	0.5508
2.554	7	2933	1.9751	0.5	98.4996	4.2960
2.1989	8.0	3352	1.8415	1.0	177.6254	5.1676
2.0739	9.0	3771	1.7894	1.0	174.7711	6.3993
1.9783	10.0	4190	1.7559	1.0	173.6515	6.9728
1.8232	11.0	4609	1.7453	1.0	173.5272	7.1491
1.7958	12.0	5028	1.7192	1.0	172.5672	7.2869
1.678	13.0	5447	1.7151	1.0	173.0770	7.1685
1.6752	14.0	5866	1.7090	1.0	174.8372	7.4453
1.551	15.0	6285	1.7027	1.0	174.9440	7.4640
1.5261	16.0	6704	1.7086	1.0	172.0875	7.4029
1.482	17.0	7123	1.7131	1.0	172.5957	7.3574
1.3927	18.0	7542	1.7207	1.0	182.4986	7.4421
1.3654	19.0	7961	1.7294	1.0	183.5361	7.5246

Framework versions

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