t5-small-finetuned-en-to-it-hrs

This model is a fine-tuned version of t5-small on the None dataset. It achieves the following results on the evaluation set:

• Loss: 2.1558
• Bleu: 9.8991
• Gen Len: 51.8287

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: 2e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 40
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Gen Len
2.0084	1.0	1125	2.8804	4.4102	67.6067
1.7918	2.0	2250	2.7757	6.1959	58.0313
1.6944	3.0	3375	2.6845	6.9152	55.6953
1.5955	4.0	4500	2.6219	7.3056	54.8213
1.5304	5.0	5625	2.5659	7.9427	53.4173
1.52	6.0	6750	2.5249	8.2049	53.678
1.4934	7.0	7875	2.4853	8.6612	52.304
1.4518	8.0	9000	2.4522	8.7991	52.6467
1.4393	9.0	10125	2.4353	8.8251	52.7047
1.4196	10.0	11250	2.4027	9.01	52.5387
1.405	11.0	12375	2.3797	9.1513	52.0273
1.3741	12.0	13500	2.3590	9.2401	52.3373
1.3693	13.0	14625	2.3378	9.3611	52.1507
1.3638	14.0	15750	2.3226	9.4213	52.2813
1.3366	15.0	16875	2.3071	9.5199	52.1507
1.3294	16.0	18000	2.2943	9.5296	51.9587
1.3258	17.0	19125	2.2788	9.6231	51.5807
1.3152	18.0	20250	2.2693	9.6586	51.8933
1.3023	19.0	21375	2.2543	9.6762	51.5733
1.3061	20.0	22500	2.2451	9.6926	51.6727
1.3004	21.0	23625	2.2344	9.773	51.6527
1.2839	22.0	24750	2.2242	9.7973	51.8113
1.2869	23.0	25875	2.2161	9.8177	51.9073
1.2819	24.0	27000	2.2115	9.8183	51.6707
1.2642	25.0	28125	2.2037	9.7645	52.0853
1.2685	26.0	29250	2.1984	9.7764	51.6927
1.2609	27.0	30375	2.1934	9.7205	51.9647
1.2585	28.0	31500	2.1834	9.8116	51.7373
1.2564	29.0	32625	2.1811	9.8547	51.8553
1.2563	30.0	33750	2.1766	9.8346	51.7293
1.258	31.0	34875	2.1748	9.8204	51.6747
1.2391	32.0	36000	2.1708	9.8485	51.7647
1.2364	33.0	37125	2.1644	9.8503	51.6713
1.2436	34.0	38250	2.1629	9.8457	51.76
1.2408	35.0	39375	2.1614	9.8899	51.6893
1.2564	36.0	40500	2.1591	9.8867	51.706
1.2318	37.0	41625	2.1575	9.866	51.782
1.2423	38.0	42750	2.1570	9.8756	51.8933
1.2399	39.0	43875	2.1558	9.8871	51.7967
1.2339	40.0	45000	2.1558	9.8991	51.8287

Framework versions

• Transformers 4.22.1
• Pytorch 1.12.1
• Datasets 2.5.1
• Tokenizers 0.11.0