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

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

• Loss: 1.7887
• Bleu: 15.4528
• Gen Len: 52.516

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
2.8637	1.0	1125	2.7212	3.496	82.846
2.6665	2.0	2250	2.5507	5.4897	65.4087
2.5307	3.0	3375	2.4286	6.688	61.9687
2.4064	4.0	4500	2.3431	7.6166	59.5613
2.3369	5.0	5625	2.2779	8.4755	57.776
2.284	6.0	6750	2.2202	9.0471	57.1227
2.2358	7.0	7875	2.1728	9.7222	55.9393
2.1747	8.0	9000	2.1357	10.4908	54.9073
2.1555	9.0	10125	2.1012	11.0378	54.292
2.1215	10.0	11250	2.0715	11.2204	54.546
2.0882	11.0	12375	2.0448	11.6557	54.1687
2.0544	12.0	13500	2.0193	12.0521	53.604
2.0355	13.0	14625	1.9959	12.2297	53.3893
2.0236	14.0	15750	1.9755	12.4706	53.3327
1.9974	15.0	16875	1.9555	12.59	53.4507
1.983	16.0	18000	1.9400	12.8305	53.1807
1.9615	17.0	19125	1.9236	13.0549	53.128
1.9519	18.0	20250	1.9111	13.1942	53.2953
1.9408	19.0	21375	1.8977	13.3979	53.332
1.9203	20.0	22500	1.8862	13.5626	52.73
1.9134	21.0	23625	1.8749	13.8549	52.904
1.8981	22.0	24750	1.8638	13.9347	53.2787
1.8911	23.0	25875	1.8557	14.1628	52.946
1.8859	24.0	27000	1.8471	14.2514	52.744
1.8692	25.0	28125	1.8406	14.4957	52.9267
1.8733	26.0	29250	1.8324	14.5489	53.112
1.8602	27.0	30375	1.8268	14.6941	52.882
1.8547	28.0	31500	1.8202	14.9101	52.948
1.8478	29.0	32625	1.8151	14.9498	52.8967
1.8485	30.0	33750	1.8102	15.0763	52.8587
1.8401	31.0	34875	1.8065	15.1604	52.8513
1.8307	32.0	36000	1.8023	15.1404	52.6533
1.8275	33.0	37125	1.7994	15.1813	52.738
1.8233	34.0	38250	1.7964	15.3185	52.7033
1.8238	35.0	39375	1.7939	15.4693	52.6433
1.8253	36.0	40500	1.7926	15.4467	52.44
1.8169	37.0	41625	1.7908	15.4167	52.5907
1.8182	38.0	42750	1.7899	15.4595	52.5433
1.8161	39.0	43875	1.7890	15.4411	52.5007
1.8169	40.0	45000	1.7887	15.4528	52.516

Framework versions

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