2c1ada40f369ac1225023c4b87b18c55

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

• Loss: 2.4506
• Data Size: 1.0
• Epoch Runtime: 64.3168
• Bleu: 4.0544

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	3.9374	0	4.5852	0.3316
No log	1	88	3.4579	0.0078	5.3821	0.4015
No log	2	176	3.1297	0.0156	9.6287	1.0607
No log	3	264	2.9770	0.0312	15.2752	1.4864
No log	4	352	2.9011	0.0625	22.4143	1.2850
No log	5	440	2.7889	0.125	24.1263	2.4666
0.2518	6	528	2.6800	0.25	28.6256	3.0062
1.036	7	616	2.5734	0.5	45.8974	3.1916
2.7437	8.0	704	2.4688	1.0	71.7194	3.4597
2.5661	9.0	792	2.4187	1.0	60.8765	3.4084
2.4016	10.0	880	2.3805	1.0	63.6305	4.0134
2.2543	11.0	968	2.3724	1.0	59.5049	3.7213
2.1267	12.0	1056	2.3690	1.0	63.6054	3.9099
2.0117	13.0	1144	2.3484	1.0	60.7380	3.7400
1.8881	14.0	1232	2.3722	1.0	62.9122	3.8564
1.7767	15.0	1320	2.3843	1.0	60.1713	4.0942
1.6562	16.0	1408	2.4124	1.0	61.9298	4.0382
1.5524	17.0	1496	2.4506	1.0	64.3168	4.0544

Framework versions

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