ab93ed2e9d34eebca5e5aa081372bebc

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

• Loss: 1.6164
• Data Size: 1.0
• Epoch Runtime: 518.7110
• Bleu: 11.2354

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	23.6601	0	39.3538	0.0188
No log	1	1286	21.2471	0.0078	43.0395	0.0080
0.4589	2	2572	10.7776	0.0156	49.1449	0.0202
0.4297	3	3858	15.8878	0.0312	58.4469	0.0380
0.4698	4	5144	2.8676	0.0625	75.4727	0.5967
2.7418	5	6430	2.1100	0.125	105.0327	8.8864
2.3992	6	7716	1.9327	0.25	164.3769	8.1645
2.1975	7	9002	1.8312	0.5	283.6717	9.0939
2.0602	8.0	10288	1.7396	1.0	517.8994	10.0316
1.903	9.0	11574	1.6808	1.0	516.5106	10.1810
1.7955	10.0	12860	1.6475	1.0	517.5944	10.6325
1.6836	11.0	14146	1.6328	1.0	517.5800	10.8419
1.6617	12.0	15432	1.6095	1.0	518.8468	10.8037
1.5574	13.0	16718	1.6049	1.0	518.3699	10.8925
1.5154	14.0	18004	1.5989	1.0	518.8894	10.8774
1.4819	15.0	19290	1.5961	1.0	519.9667	11.0827
1.4005	16.0	20576	1.6002	1.0	522.3767	11.1317
1.3689	17.0	21862	1.6039	1.0	519.4761	11.1058
1.3182	18.0	23148	1.6082	1.0	519.4104	11.0625
1.2775	19.0	24434	1.6164	1.0	518.7110	11.2354

Framework versions

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