b681866e14ce1df397f72335e0e0272e

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

• Loss: 1.5306
• Data Size: 1.0
• Epoch Runtime: 216.8754
• Bleu: 6.5167

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	3.4089	0	16.6455	0.2551
No log	1	390	2.6357	0.0078	18.5474	1.1720
No log	2	780	2.4830	0.0156	24.8775	1.2194
No log	3	1170	2.3850	0.0312	31.0326	1.4041
No log	4	1560	2.2682	0.0625	40.1484	1.8537
0.1567	5	1950	2.1291	0.125	52.7841	2.3315
0.3635	6	2340	1.9819	0.25	73.9581	3.1243
2.1355	7	2730	1.8350	0.5	123.9935	3.7571
1.885	8.0	3120	1.6774	1.0	217.7282	4.5244
1.6935	9.0	3510	1.5903	1.0	219.2729	5.0980
1.5639	10.0	3900	1.5375	1.0	217.0207	5.5020
1.4526	11.0	4290	1.4975	1.0	215.9094	5.7430
1.345	12.0	4680	1.4803	1.0	218.5554	5.9550
1.2678	13.0	5070	1.4667	1.0	220.3127	6.1438
1.1842	14.0	5460	1.4640	1.0	216.3124	6.2584
1.082	15.0	5850	1.4679	1.0	215.7120	6.3342
1.0072	16.0	6240	1.4903	1.0	216.9731	6.4530
0.9402	17.0	6630	1.4902	1.0	217.8165	6.4267
0.8681	18.0	7020	1.5306	1.0	216.8754	6.5167

Framework versions

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