a321e9b0659813244a5ad1d93b520f34

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

• Loss: 2.6203
• Data Size: 1.0
• Epoch Runtime: 15.6628
• Bleu: 3.4827

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	14.9715	0	2.3061	0.0018
No log	1	58	15.0087	0.0078	3.2960	0.0015
No log	2	116	14.6568	0.0156	2.4653	0.0016
No log	3	174	14.2637	0.0312	3.4635	0.0015
No log	4	232	13.2767	0.0625	3.8368	0.0022
No log	5	290	11.9184	0.125	4.9878	0.0019
1.3411	6	348	10.1487	0.25	6.9704	0.0030
1.7821	7	406	6.9995	0.5	10.1543	0.0097
5.9912	8.0	464	3.9077	1.0	16.8886	0.1249
4.7957	9.0	522	3.4031	1.0	14.4762	0.2553
4.2267	10.0	580	3.2038	1.0	15.1728	1.1483
3.9956	11.0	638	3.0814	1.0	14.4346	1.6017
3.779	12.0	696	2.9809	1.0	14.9137	1.7514
3.5176	13.0	754	2.9191	1.0	14.8790	1.9389
3.4353	14.0	812	2.8774	1.0	15.9237	2.0958
3.3455	15.0	870	2.8332	1.0	14.4657	2.2009
3.2709	16.0	928	2.8029	1.0	15.2881	2.3801
3.2204	17.0	986	2.7743	1.0	15.8871	2.5094
3.1535	18.0	1044	2.7560	1.0	15.8395	2.6264
3.0708	19.0	1102	2.7394	1.0	16.7076	2.6227
3.0073	20.0	1160	2.7232	1.0	14.6159	2.7494
2.9694	21.0	1218	2.7039	1.0	14.4745	2.7429
2.9389	22.0	1276	2.6909	1.0	15.2147	2.9550
2.9015	23.0	1334	2.6829	1.0	14.9436	2.9738
2.8674	24.0	1392	2.6747	1.0	15.5361	3.0137
2.8004	25.0	1450	2.6644	1.0	16.3334	3.0425
2.7664	26.0	1508	2.6567	1.0	14.9050	3.0071
2.7162	27.0	1566	2.6462	1.0	15.3831	3.1339
2.7098	28.0	1624	2.6492	1.0	15.9688	3.1105
2.693	29.0	1682	2.6431	1.0	16.7174	3.2350
2.65	30.0	1740	2.6424	1.0	17.3541	3.1945
2.6337	31.0	1798	2.6336	1.0	17.5802	3.2337
2.5747	32.0	1856	2.6273	1.0	14.9585	3.3381
2.5469	33.0	1914	2.6282	1.0	15.5778	3.2941
2.5217	34.0	1972	2.6253	1.0	15.5397	3.2940
2.5028	35.0	2030	2.6194	1.0	16.0798	3.3374
2.47	36.0	2088	2.6203	1.0	16.1742	3.3892
2.4529	37.0	2146	2.6243	1.0	17.2893	3.3439
2.4071	38.0	2204	2.6181	1.0	15.5731	3.3623
2.3768	39.0	2262	2.6158	1.0	15.6917	3.4376
2.3709	40.0	2320	2.6265	1.0	15.6787	3.4468
2.334	41.0	2378	2.6278	1.0	16.2384	3.4655
2.3198	42.0	2436	2.6287	1.0	16.2185	3.4571
2.2941	43.0	2494	2.6203	1.0	15.6628	3.4827

Framework versions

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