t5-small-finetuned-en-to-it-lrs

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

• Loss: 2.1483
• Bleu: 10.4962
• Gen Len: 51.8247

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
1.9618	1.0	1125	2.8717	4.6688	66.512
1.7256	2.0	2250	2.7638	6.5673	56.7267
1.6133	3.0	3375	2.6703	7.4218	55.1753
1.5132	4.0	4500	2.6096	7.9581	54.5387
1.4558	5.0	5625	2.5603	8.5191	52.41
1.4392	6.0	6750	2.5109	8.976	52.1867
1.4113	7.0	7875	2.4768	9.2615	51.8907
1.3669	8.0	9000	2.4447	9.3001	52.6
1.3575	9.0	10125	2.4262	9.5818	51.774
1.3315	10.0	11250	2.3906	9.584	52.3213
1.3231	11.0	12375	2.3740	9.7574	51.63
1.2917	12.0	13500	2.3475	9.8298	51.6367
1.282	13.0	14625	2.3269	9.8176	52.06
1.2841	14.0	15750	2.3121	9.9668	51.9487
1.2548	15.0	16875	2.2993	9.9941	51.708
1.2487	16.0	18000	2.2816	10.0288	52.364
1.2462	17.0	19125	2.2697	10.1991	51.3893
1.232	18.0	20250	2.2581	10.2667	51.6693
1.2227	19.0	21375	2.2428	10.3357	51.5373
1.2279	20.0	22500	2.2350	10.3646	51.4633
1.2159	21.0	23625	2.2275	10.3489	51.472
1.2036	22.0	24750	2.2186	10.3756	51.444
1.2089	23.0	25875	2.2082	10.3555	51.7133
1.1957	24.0	27000	2.2016	10.4624	51.4293
1.1828	25.0	28125	2.1953	10.4474	51.4287
1.1885	26.0	29250	2.1887	10.3417	51.4227
1.1817	27.0	30375	2.1844	10.4777	51.5787
1.1769	28.0	31500	2.1759	10.4044	51.5907
1.1831	29.0	32625	2.1728	10.4434	51.6587
1.1842	30.0	33750	2.1706	10.4136	51.7653
1.1828	31.0	34875	2.1689	10.5099	51.5893
1.1673	32.0	36000	2.1613	10.4957	51.646
1.1603	33.0	37125	2.1570	10.4438	51.6633
1.1718	34.0	38250	2.1564	10.5364	51.7113
1.1651	35.0	39375	2.1538	10.4444	51.6593
1.1756	36.0	40500	2.1501	10.4497	51.7393
1.1595	37.0	41625	2.1499	10.4701	51.7313
1.1603	38.0	42750	2.1499	10.4611	51.7533
1.1586	39.0	43875	2.1487	10.4776	51.836
1.161	40.0	45000	2.1483	10.4962	51.8247

Framework versions

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