f9237b85afe320e346c08794a374f118

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

• Loss: 1.0682
• Data Size: 1.0
• Epoch Runtime: 420.0688
• Bleu: 8.8443

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	3.3130	0	27.7994	0.5982
No log	1	1000	2.7566	0.0078	36.8077	1.6539
No log	2	2000	2.5586	0.0156	38.8126	1.7336
No log	3	3000	2.3754	0.0312	45.6193	2.1060
0.0935	4	4000	2.2032	0.0625	58.2706	2.7375
2.3525	5	5000	2.0270	0.125	84.7273	3.2737
0.1298	6	6000	1.8329	0.25	130.1406	4.0081
0.1659	7	7000	1.6288	0.5	230.0070	4.9057
1.5884	8.0	8000	1.4209	1.0	425.6913	6.1489
1.4317	9.0	9000	1.3147	1.0	435.2061	6.7954
1.3169	10.0	10000	1.2400	1.0	426.8136	7.2651
1.2384	11.0	11000	1.1938	1.0	447.4447	7.6365
1.1637	12.0	12000	1.1566	1.0	444.3004	7.8984
1.1049	13.0	13000	1.1355	1.0	429.0359	8.0949
1.0587	14.0	14000	1.1126	1.0	429.8473	8.1326
0.997	15.0	15000	1.1031	1.0	422.7297	8.2998
0.9601	16.0	16000	1.0837	1.0	428.8351	8.4219
0.8953	17.0	17000	1.0751	1.0	432.1069	8.4233
0.888	18.0	18000	1.0752	1.0	428.3797	8.4970
0.8423	19.0	19000	1.0645	1.0	414.9656	8.6346
0.799	20.0	20000	1.0649	1.0	429.7402	8.6474
0.7798	21.0	21000	1.0522	1.0	445.6019	8.6459
0.7559	22.0	22000	1.0573	1.0	427.1178	8.7030
0.7149	23.0	23000	1.0628	1.0	434.0112	8.7722
0.6874	24.0	24000	1.0620	1.0	426.1050	8.7406
0.6829	25.0	25000	1.0682	1.0	420.0688	8.8443

Framework versions

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