be37dcb56e0edd7abfdd701c1e6cf0df

This model is a fine-tuned version of facebook/mbart-large-50 on the Helsinki-NLP/opus_books [de-en] dataset. It achieves the following results on the evaluation set:

• Loss: 2.3852
• Data Size: 1.0
• Epoch Runtime: 322.3205
• Bleu: 9.3174

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	6.2567	0	26.7048	1.6973
No log	1	1286	4.4748	0.0078	29.5383	5.0592
0.0895	2	2572	3.3679	0.0156	33.2817	10.4733
0.0899	3	3858	2.2590	0.0312	37.5992	6.7743
0.0993	4	5144	2.1548	0.0625	47.5234	8.0927
2.0738	5	6430	2.0462	0.125	66.0811	8.5674
1.9511	6	7716	1.9638	0.25	101.6687	11.8428
1.8928	7	9002	1.9670	0.5	176.5597	12.5154
1.8482	8.0	10288	1.8666	1.0	321.5913	9.3952
1.4755	9.0	11574	1.8589	1.0	322.5186	9.5484
1.2218	10.0	12860	1.9179	1.0	322.5333	9.4680
0.9689	11.0	14146	2.0313	1.0	322.2198	9.4077
0.8023	12.0	15432	2.2226	1.0	324.6607	9.6218
0.6054	13.0	16718	2.3852	1.0	322.3205	9.3174

Framework versions

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