55cb0a9cd40a9a3d96abe02117faa2c7

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

• Loss: 2.1927
• Data Size: 1.0
• Epoch Runtime: 117.2320
• Bleu: 5.6841

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	26.7755	0	10.5743	0.0035
No log	1	806	23.7421	0.0078	12.4043	0.0038
No log	2	1612	19.8363	0.0156	12.5261	0.0038
No log	3	2418	14.3395	0.0312	14.7295	0.0043
0.6163	4	3224	10.0032	0.0625	17.9058	0.0080
8.8754	5	4030	5.8672	0.125	24.5137	0.0109
5.1782	6	4836	3.8117	0.25	38.0788	0.3905
4.357	7	5642	3.4257	0.5	64.5593	0.9788
3.9119	8.0	6448	3.1506	1.0	118.7890	1.6468
3.6699	9.0	7254	3.0027	1.0	119.7912	2.1614
3.5151	10.0	8060	2.8978	1.0	116.5340	2.5095
3.4398	11.0	8866	2.8258	1.0	115.2833	2.7432
3.3133	12.0	9672	2.7666	1.0	115.3615	2.9873
3.1923	13.0	10478	2.7178	1.0	113.9382	3.1817
3.1757	14.0	11284	2.6739	1.0	117.3042	3.2654
3.0818	15.0	12090	2.6356	1.0	116.9318	3.4776
3.0441	16.0	12896	2.6015	1.0	117.2625	3.5692
3.0108	17.0	13702	2.5726	1.0	117.9902	3.6967
2.8977	18.0	14508	2.5473	1.0	116.5249	3.8217
2.92	19.0	15314	2.5208	1.0	117.9966	3.9321
2.8672	20.0	16120	2.4957	1.0	115.1913	4.0253
2.8529	21.0	16926	2.4786	1.0	117.6613	4.1228
2.8122	22.0	17732	2.4582	1.0	116.3007	4.2133
2.7926	23.0	18538	2.4428	1.0	118.0086	4.2989
2.7573	24.0	19344	2.4212	1.0	119.3303	4.3737
2.6993	25.0	20150	2.4048	1.0	121.9447	4.4787
2.6851	26.0	20956	2.3906	1.0	121.3424	4.5364
2.6485	27.0	21762	2.3807	1.0	122.2796	4.6086
2.6095	28.0	22568	2.3649	1.0	119.1250	4.6811
2.5946	29.0	23374	2.3576	1.0	117.5540	4.7321
2.5722	30.0	24180	2.3418	1.0	117.3148	4.7818
2.5593	31.0	24986	2.3239	1.0	118.0510	4.8432
2.5273	32.0	25792	2.3174	1.0	117.3594	4.8867
2.5222	33.0	26598	2.3045	1.0	117.6567	4.9687
2.4879	34.0	27404	2.2970	1.0	117.5826	5.0252
2.4911	35.0	28210	2.2897	1.0	116.6725	5.0981
2.4489	36.0	29016	2.2803	1.0	117.1047	5.1194
2.4286	37.0	29822	2.2721	1.0	117.3116	5.2134
2.4349	38.0	30628	2.2595	1.0	116.3583	5.2308
2.4397	39.0	31434	2.2546	1.0	116.7340	5.2712
2.373	40.0	32240	2.2515	1.0	117.0596	5.3281
2.4112	41.0	33046	2.2414	1.0	117.8659	5.3466
2.3531	42.0	33852	2.2321	1.0	118.6556	5.4046
2.3229	43.0	34658	2.2268	1.0	117.8467	5.4281
2.2869	44.0	35464	2.2250	1.0	116.1971	5.5117
2.3362	45.0	36270	2.2130	1.0	116.3434	5.5182
2.3043	46.0	37076	2.2117	1.0	116.3559	5.5532
2.2528	47.0	37882	2.2042	1.0	116.1060	5.5854
2.2729	48.0	38688	2.1900	1.0	117.0202	5.6367
2.2545	49.0	39494	2.1979	1.0	116.4764	5.6790
2.2264	50.0	40300	2.1927	1.0	117.2320	5.6841

Framework versions

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