30aa6304118bcbb5b00adcc6a36673af

This model is a fine-tuned version of google-t5/t5-3b on the Helsinki-NLP/opus_books [de-fr] dataset. It achieves the following results on the evaluation set:

• Loss: 1.3122
• Data Size: 1.0
• Epoch Runtime: 397.1098
• Bleu: 10.3923

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	2.0334	0	25.0784	4.9579
No log	1	872	1.7639	0.0078	29.5733	6.7638
No log	2	1744	1.6814	0.0156	38.5627	7.8419
0.0295	3	2616	1.6170	0.0312	49.5605	7.2458
0.1116	4	3488	1.5600	0.0625	66.5870	7.7329
1.6824	5	4360	1.4948	0.125	81.7043	8.0213
1.5638	6	5232	1.4233	0.25	138.3493	8.8669
1.4028	7	6104	1.3389	0.5	228.3327	9.5633
1.2819	8.0	6976	1.2618	1.0	420.3590	10.2029
1.1336	9.0	7848	1.2301	1.0	416.9507	10.5613
1.0311	10.0	8720	1.2166	1.0	416.8892	10.4853
0.9222	11.0	9592	1.2308	1.0	411.2438	10.4009
0.811	12.0	10464	1.2448	1.0	413.9808	10.6733
0.7448	13.0	11336	1.2777	1.0	415.1636	10.4867
0.6513	14.0	12208	1.3122	1.0	397.1098	10.3923

Framework versions

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