c5b2b95ea9abec1d81d1d74441cf49bb

This model is a fine-tuned version of google/mt5-xl on the Helsinki-NLP/opus_books [de-en] dataset. It achieves the following results on the evaluation set:

• Loss: 1.5675
• Data Size: 1.0
• Epoch Runtime: 596.5196
• Bleu: 11.8990

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	7.3516	0	38.9813	0.0429
No log	1	1286	3.1170	0.0078	44.0128	0.8280
0.0842	2	2572	2.2121	0.0156	54.4514	1.2466
0.0721	3	3858	1.9391	0.0312	66.8970	7.0472
0.0981	4	5144	1.8078	0.0625	85.2423	8.7323
2.0268	5	6430	1.7242	0.125	118.0626	10.8961
1.9028	6	7716	1.6520	0.25	187.9390	10.9875
1.7664	7	9002	1.5789	0.5	323.0581	10.4080
1.6369	8.0	10288	1.5044	1.0	595.9808	11.3215
1.4505	9.0	11574	1.4778	1.0	596.4580	11.4040
1.286	10.0	12860	1.4823	1.0	593.6933	11.5654
1.1518	11.0	14146	1.4947	1.0	593.1440	11.6962
1.0809	12.0	15432	1.5262	1.0	595.4774	11.8040
0.9338	13.0	16718	1.5675	1.0	596.5196	11.8990

Framework versions

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