97be97fd64b44194fa4285ea3d731c6f

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

• Loss: 1.8195
• Data Size: 1.0
• Epoch Runtime: 270.7832
• Bleu: 9.9824

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	15.0430	0	22.1078	0.0048
No log	1	1286	11.7384	0.0078	25.0985	0.0088
0.2772	2	2572	8.8585	0.0156	27.4474	0.0104
0.2937	3	3858	5.6575	0.0312	31.5020	0.0216
0.2958	4	5144	3.1860	0.0625	39.5076	1.0036
3.686	5	6430	2.7062	0.125	55.7782	4.8580
3.1186	6	7716	2.4233	0.25	87.3212	4.9989
2.821	7	9002	2.2720	0.5	150.8888	5.6347
2.5885	8.0	10288	2.1160	1.0	278.0043	7.5789
2.4115	9.0	11574	2.0445	1.0	269.7820	7.9766
2.2858	10.0	12860	1.9866	1.0	269.2880	8.4875
2.1837	11.0	14146	1.9571	1.0	270.2258	8.7403
2.1721	12.0	15432	1.9208	1.0	267.5333	8.8313
2.063	13.0	16718	1.9012	1.0	268.5411	9.0129
2.0218	14.0	18004	1.8844	1.0	266.7837	8.9778
2.0103	15.0	19290	1.8685	1.0	267.9107	9.0768
1.9174	16.0	20576	1.8555	1.0	267.7329	9.1727
1.898	17.0	21862	1.8460	1.0	267.6506	9.1968
1.8506	18.0	23148	1.8421	1.0	266.0922	9.2619
1.8408	19.0	24434	1.8290	1.0	266.7357	9.4649
1.7599	20.0	25720	1.8216	1.0	266.7927	9.4724
1.7454	21.0	27006	1.8163	1.0	267.5041	9.4511
1.732	22.0	28292	1.8134	1.0	267.9240	9.6085
1.7021	23.0	29578	1.8097	1.0	267.9228	9.7375
1.6528	24.0	30864	1.8143	1.0	266.7621	9.7147
1.6171	25.0	32150	1.8100	1.0	266.2737	9.6732
1.6328	26.0	33436	1.8043	1.0	269.0423	9.8114
1.5665	27.0	34722	1.8085	1.0	266.6011	9.8003
1.5345	28.0	36008	1.8146	1.0	265.0006	9.8266
1.5288	29.0	37294	1.8034	1.0	267.2608	9.8642
1.4943	30.0	38580	1.8111	1.0	267.3260	9.8983
1.4335	31.0	39866	1.8179	1.0	267.3371	9.9456
1.4141	32.0	41152	1.8110	1.0	267.1492	9.9939
1.4011	33.0	42438	1.8195	1.0	270.7832	9.9824

Framework versions

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