t5-base_ro-finetuned-en-to-it

This model is a fine-tuned version of j0hngou/t5-base-finetuned-en-to-ro on the ccmatrix dataset. It achieves the following results on the evaluation set:

• Loss: 1.4669
• Bleu: 19.6396
• Gen Len: 52.4247

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
No log	1.0	282	2.0942	5.6875	73.434
2.5108	2.0	564	1.9725	6.6631	72.6607
2.5108	3.0	846	1.9010	7.9227	67.01
2.1659	4.0	1128	1.8452	8.9935	65.1027
2.1659	5.0	1410	1.7979	9.4164	64.9827
2.0288	6.0	1692	1.7590	9.6035	66.6933
2.0288	7.0	1974	1.7264	10.7658	62.068
1.9238	8.0	2256	1.6955	11.5779	59.472
1.8435	9.0	2538	1.6729	12.7588	56.584
1.8435	10.0	2820	1.6541	13.3086	56.1153
1.775	11.0	3102	1.6337	13.8543	55.3307
1.775	12.0	3384	1.6148	14.3566	55.2853
1.7204	13.0	3666	1.5994	14.693	55.6607
1.7204	14.0	3948	1.5838	15.1284	55.5327
1.6705	15.0	4230	1.5742	15.6125	55.0087
1.632	16.0	4512	1.5600	15.9616	54.052
1.632	17.0	4794	1.5526	16.495	53.562
1.5868	18.0	5076	1.5392	16.4252	54.4613
1.5868	19.0	5358	1.5311	16.753	54.1853
1.5656	20.0	5640	1.5262	17.0308	54.2473
1.5656	21.0	5922	1.5186	17.3553	53.396
1.529	22.0	6204	1.5121	17.6177	53.472
1.529	23.0	6486	1.5058	17.6409	53.6847
1.5071	24.0	6768	1.5038	18.2009	53.2327
1.4903	25.0	7050	1.4962	18.4838	52.9587
1.4903	26.0	7332	1.4935	18.5545	52.688
1.4686	27.0	7614	1.4879	18.62	53.5
1.4686	28.0	7896	1.4850	19.0099	52.34
1.4511	29.0	8178	1.4813	19.0538	52.474
1.4511	30.0	8460	1.4787	18.89	53.0753
1.4364	31.0	8742	1.4756	19.2582	52.3587
1.4279	32.0	9024	1.4739	19.2973	52.69
1.4279	33.0	9306	1.4725	19.3624	52.694
1.4172	34.0	9588	1.4704	19.5421	52.1667
1.4172	35.0	9870	1.4689	19.4807	52.5487
1.4141	36.0	10152	1.4685	19.5972	52.2733
1.4141	37.0	10434	1.4676	19.5835	52.374
1.4058	38.0	10716	1.4674	19.6374	52.3447
1.4058	39.0	10998	1.4671	19.6105	52.5273
1.4027	40.0	11280	1.4669	19.6396	52.4247

Framework versions

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