f7a9c47619734b98bc7b8fa773b4e726

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

• Loss: 1.0939
• Data Size: 1.0
• Epoch Runtime: 546.1156
• Bleu: 7.9002

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	3.4278	0	40.9774	0.2095
No log	1	1000	2.5764	0.0078	45.0469	0.9001
No log	2	2000	2.3949	0.0156	53.7006	1.0212
No log	3	3000	2.2016	0.0312	65.0629	1.3964
0.0964	4	4000	2.0627	0.0625	81.2244	1.8657
2.3311	5	5000	1.9021	0.125	109.4563	2.4578
0.1257	6	6000	1.7099	0.25	172.1917	3.2116
0.1584	7	7000	1.5225	0.5	293.2010	4.1032
1.4893	8.0	8000	1.3268	1.0	538.2769	5.2577
1.311	9.0	9000	1.2197	1.0	543.0087	6.0068
1.1839	10.0	10000	1.1529	1.0	535.1504	6.5436
1.0855	11.0	11000	1.1144	1.0	544.5768	6.8595
0.9977	12.0	12000	1.0864	1.0	539.8519	7.1574
0.9255	13.0	13000	1.0694	1.0	540.7354	7.4193
0.8519	14.0	14000	1.0626	1.0	537.0915	7.5882
0.7797	15.0	15000	1.0628	1.0	548.4139	7.7185
0.7195	16.0	16000	1.0635	1.0	538.1533	7.7641
0.6479	17.0	17000	1.0742	1.0	543.9543	7.8529
0.6135	18.0	18000	1.0939	1.0	546.1156	7.9002

Framework versions

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