din0s/t5-small-fr-finetuned-en-to-it

t5-small_fr-finetuned-en-to-it

This model is a fine-tuned version of din0s/t5-small-finetuned-en-to-fr on the ccmatrix dataset. It achieves the following results on the evaluation set:

• Loss: 2.3225
• Bleu: 7.4222
• Gen Len: 59.1127

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: 96
• eval_batch_size: 96
• 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	94	3.0406	3.2546	52.6127
No log	2.0	188	2.9278	3.1206	62.774
No log	3.0	282	2.8573	3.4206	63.6707
No log	4.0	376	2.8030	3.4847	66.408
No log	5.0	470	2.7602	3.8933	64.362
3.2982	6.0	564	2.7185	3.9298	66.058
3.2982	7.0	658	2.6842	4.0344	65.5773
3.2982	8.0	752	2.6536	4.3243	65.0047
3.2982	9.0	846	2.6233	4.5078	64.5813
3.2982	10.0	940	2.5966	4.6657	63.654
2.9837	11.0	1034	2.5743	4.7664	63.326
2.9837	12.0	1128	2.5526	4.9535	62.7327
2.9837	13.0	1222	2.5303	5.1386	63.5887
2.9837	14.0	1316	2.5122	5.1037	64.1667
2.9837	15.0	1410	2.4937	5.3304	63.116
2.8416	16.0	1504	2.4797	5.5006	61.4953
2.8416	17.0	1598	2.4627	5.5892	62.01
2.8416	18.0	1692	2.4497	5.8497	61.42
2.8416	19.0	1786	2.4372	6.0074	61.1587
2.8416	20.0	1880	2.4256	6.1464	60.522
2.8416	21.0	1974	2.4148	6.3117	59.5567
2.7428	22.0	2068	2.4039	6.4626	59.532
2.7428	23.0	2162	2.3939	6.5287	60.2307
2.7428	24.0	2256	2.3857	6.6093	60.22
2.7428	25.0	2350	2.3772	6.8004	59.396
2.7428	26.0	2444	2.3703	6.9433	59.5027
2.6779	27.0	2538	2.3631	7.0153	59.1433
2.6779	28.0	2632	2.3575	7.1783	58.9793
2.6779	29.0	2726	2.3514	7.1639	59.362
2.6779	30.0	2820	2.3457	7.2176	58.9927
2.6779	31.0	2914	2.3411	7.2599	59.1433
2.6335	32.0	3008	2.3374	7.284	59.1787
2.6335	33.0	3102	2.3339	7.3678	59.07
2.6335	34.0	3196	2.3307	7.3364	58.9813
2.6335	35.0	3290	2.3281	7.3318	58.96
2.6335	36.0	3384	2.3259	7.394	59.0787
2.6335	37.0	3478	2.3245	7.4133	59.0393
2.609	38.0	3572	2.3232	7.383	59.1887
2.609	39.0	3666	2.3227	7.4105	59.1227
2.609	40.0	3760	2.3225	7.4222	59.1127

Framework versions

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