3093fb9a8983489518e12dfaa044f155

This model is a fine-tuned version of google/mt5-base on the Helsinki-NLP/opus_books [de-fr] dataset. It achieves the following results on the evaluation set:

• Loss: 1.5731
• Data Size: 1.0
• Epoch Runtime: 188.3161
• Bleu: 8.7305

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: 5e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• total_train_batch_size: 32
• total_eval_batch_size: 32
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: constant
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Data Size	Epoch Runtime	Bleu
No log	0	0	14.3450	0	15.1002	0.0066
No log	1	872	12.6367	0.0078	17.3453	0.0071
No log	2	1744	11.3200	0.0156	19.4900	0.0055
0.2313	3	2616	9.8811	0.0312	21.5632	0.0055
0.7249	4	3488	7.3901	0.0625	26.8591	0.0050
7.727	5	4360	4.0420	0.125	37.4920	0.0342
3.4298	6	5232	2.3942	0.25	61.2757	2.2865
2.8198	7	6104	2.1322	0.5	102.7155	3.3724
2.5025	8.0	6976	1.9606	1.0	188.6247	4.6341
2.3319	9.0	7848	1.8750	1.0	189.2860	5.6087
2.2251	10.0	8720	1.8240	1.0	185.6852	6.1172
2.1327	11.0	9592	1.7806	1.0	185.8606	6.6380
2.0174	12.0	10464	1.7439	1.0	185.1414	6.9475
1.9702	13.0	11336	1.7121	1.0	185.1029	7.2153
1.9087	14.0	12208	1.7008	1.0	187.1098	7.2857
1.8636	15.0	13080	1.6774	1.0	186.6470	7.5258
1.8069	16.0	13952	1.6552	1.0	185.0405	7.7200
1.7322	17.0	14824	1.6388	1.0	185.7621	7.8547
1.7076	18.0	15696	1.6320	1.0	185.3292	8.0456
1.6858	19.0	16568	1.6202	1.0	186.4298	8.1104
1.6269	20.0	17440	1.6147	1.0	187.6369	8.1988
1.6425	21.0	18312	1.6047	1.0	184.5107	8.2656
1.5834	22.0	19184	1.5977	1.0	184.2165	8.2792
1.5241	23.0	20056	1.5924	1.0	185.0867	8.2898
1.5079	24.0	20928	1.5914	1.0	186.8553	8.4439
1.4828	25.0	21800	1.5870	1.0	186.9477	8.5025
1.4705	26.0	22672	1.5778	1.0	184.4874	8.5475
1.4329	27.0	23544	1.5781	1.0	184.5486	8.5273
1.3983	28.0	24416	1.5818	1.0	184.5743	8.6452
1.36	29.0	25288	1.5694	1.0	185.7436	8.6280
1.3852	30.0	26160	1.5794	1.0	186.8681	8.6158
1.3502	31.0	27032	1.5800	1.0	185.2266	8.6586
1.3035	32.0	27904	1.5788	1.0	185.1830	8.7157
1.3	33.0	28776	1.5731	1.0	188.3161	8.7305

Framework versions

• Transformers 4.57.0
• Pytorch 2.8.0+cu128
• Datasets 4.2.0
• Tokenizers 0.22.1