t5-base_fr-finetuned-en-to-it

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

• Loss: 1.4677
• Bleu: 20.3152
• Gen Len: 51.4433

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	2.0344	6.8826	64.574
2.3997	2.0	564	1.9371	7.9377	64.274
2.3997	3.0	846	1.8740	9.2364	59.8673
2.145	4.0	1128	1.8240	9.8068	60.566
2.145	5.0	1410	1.7813	10.3961	60.106
2.0183	6.0	1692	1.7476	11.2005	59.032
2.0183	7.0	1974	1.7152	11.8127	58.1673
1.9185	8.0	2256	1.6872	12.4843	57.5787
1.8414	9.0	2538	1.6643	13.4338	55.502
1.8414	10.0	2820	1.6459	13.7847	55.6753
1.7755	11.0	3102	1.6273	14.6959	53.838
1.7755	12.0	3384	1.6121	15.2948	53.4127
1.7224	13.0	3666	1.5967	15.878	53.0733
1.7224	14.0	3948	1.5809	16.3788	52.778
1.6751	15.0	4230	1.5689	16.7415	52.8
1.6358	16.0	4512	1.5580	17.0318	52.854
1.6358	17.0	4794	1.5509	17.6302	52.0947
1.5921	18.0	5076	1.5389	17.4239	52.71
1.5921	19.0	5358	1.5317	17.9003	52.3427
1.5696	20.0	5640	1.5253	17.769	52.928
1.5696	21.0	5922	1.5172	18.2974	51.8173
1.5344	22.0	6204	1.5117	18.5755	52.012
1.5344	23.0	6486	1.5046	18.5362	52.1447
1.5136	24.0	6768	1.5034	18.7394	51.9887
1.4968	25.0	7050	1.4968	19.1622	51.736
1.4968	26.0	7332	1.4947	19.1636	51.8467
1.472	27.0	7614	1.4886	19.3845	51.774
1.472	28.0	7896	1.4844	19.5481	51.458
1.4575	29.0	8178	1.4827	19.739	51.4593
1.4575	30.0	8460	1.4791	19.818	51.62
1.4435	31.0	8742	1.4763	19.904	51.5167
1.4336	32.0	9024	1.4750	19.9507	51.3787
1.4336	33.0	9306	1.4742	20.0704	51.3527
1.4236	34.0	9588	1.4717	20.2553	51.1967
1.4236	35.0	9870	1.4705	20.3014	51.156
1.4188	36.0	10152	1.4697	20.2263	51.4173
1.4188	37.0	10434	1.4687	20.244	51.394
1.412	38.0	10716	1.4681	20.2699	51.5993
1.412	39.0	10998	1.4676	20.2758	51.4473
1.4087	40.0	11280	1.4677	20.3152	51.4433

Framework versions

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