db020be9e4e35d83e65d19fab946b815

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

• Loss: 1.8069
• Data Size: 1.0
• Epoch Runtime: 180.3731
• Bleu: 6.5813

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	17.6779	0	14.1149	0.0142
No log	1	808	15.1169	0.0078	16.4196	0.0167
No log	2	1616	12.7227	0.0156	17.2645	0.0185
No log	3	2424	9.3009	0.0312	20.6988	0.0207
0.3854	4	3232	6.4219	0.0625	25.8235	0.0170
5.5513	5	4040	2.9263	0.125	35.1693	0.5854
3.519	6	4848	2.5624	0.25	55.7010	2.1025
3.0277	7	5656	2.3566	0.5	95.1281	2.8146
2.7929	8.0	6464	2.1938	1.0	177.9741	3.7082
2.5748	9.0	7272	2.1051	1.0	172.6640	4.2122
2.4543	10.0	8080	2.0541	1.0	171.3369	4.6003
2.3596	11.0	8888	2.0055	1.0	173.2188	4.9057
2.2539	12.0	9696	1.9744	1.0	173.8705	5.1304
2.2171	13.0	10504	1.9486	1.0	172.1368	5.2480
2.1552	14.0	11312	1.9204	1.0	171.5559	5.5457
2.1167	15.0	12120	1.9034	1.0	174.3264	5.6359
2.0847	16.0	12928	1.8819	1.0	172.6913	5.8096
2.0243	17.0	13736	1.8685	1.0	171.9891	5.8948
1.9705	18.0	14544	1.8581	1.0	173.3638	5.9111
1.9269	19.0	15352	1.8497	1.0	174.2880	6.0771
1.889	20.0	16160	1.8420	1.0	172.6429	6.1777
1.8598	21.0	16968	1.8348	1.0	176.2967	6.2359
1.8036	22.0	17776	1.8234	1.0	179.8974	6.2606
1.7798	23.0	18584	1.8137	1.0	180.2397	6.3606
1.7423	24.0	19392	1.8092	1.0	181.6130	6.4088
1.7068	25.0	20200	1.8075	1.0	180.5901	6.4412
1.6844	26.0	21008	1.8093	1.0	181.9296	6.5005
1.6669	27.0	21816	1.8051	1.0	181.5158	6.4800
1.6396	28.0	22624	1.7990	1.0	180.2235	6.5655
1.6157	29.0	23432	1.8001	1.0	181.6333	6.5861
1.549	30.0	24240	1.8034	1.0	180.0172	6.6056
1.5684	31.0	25048	1.7994	1.0	179.4207	6.6328
1.5378	32.0	25856	1.8069	1.0	180.3731	6.5813

Framework versions

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