4207412ef4d8bdb4097e0fa03b7a5dc0

This model is a fine-tuned version of google/long-t5-tglobal-xl on the Helsinki-NLP/opus_books [fr-no] dataset. It achieves the following results on the evaluation set:

• Loss: 1.8462
• Data Size: 1.0
• Epoch Runtime: 60.4626
• Bleu: 2.9068

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	3.6425	0	4.2518	0.0742
No log	1	86	3.0854	0.0078	5.1244	0.1341
No log	2	172	2.7658	0.0156	9.0773	0.5211
No log	3	258	2.6274	0.0312	14.8940	0.6222
No log	4	344	2.5109	0.0625	20.4329	0.5751
0.1673	5	430	2.3721	0.125	23.8524	0.6793
0.6914	6	516	2.2496	0.25	28.4925	0.8805
0.9068	7	602	2.1193	0.5	37.6588	1.2060
1.3423	8.0	688	1.9904	1.0	64.7445	1.6440
2.0956	9.0	774	1.9233	1.0	59.7234	1.8960
1.956	10.0	860	1.8783	1.0	58.2873	2.2163
1.8623	11.0	946	1.8414	1.0	63.2406	2.2306
1.7169	12.0	1032	1.8313	1.0	57.9542	2.3094
1.6203	13.0	1118	1.8149	1.0	58.2547	2.5097
1.5128	14.0	1204	1.7892	1.0	62.9974	2.5785
1.4183	15.0	1290	1.7974	1.0	58.1925	2.6963
1.3223	16.0	1376	1.8240	1.0	62.8230	2.7326
1.2298	17.0	1462	1.8259	1.0	58.4253	2.8951
1.1489	18.0	1548	1.8462	1.0	60.4626	2.9068

Framework versions

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