d97211242dece02bd8b2d809a3a38bf1

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

• Loss: 1.0666
• Data Size: 1.0
• Epoch Runtime: 736.0040
• Bleu: 15.4028

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	1.6684	0	74.0586	10.0814
No log	1	3177	1.5800	0.0078	68.1763	10.9074
0.027	2	6354	1.5189	0.0156	76.0990	11.3439
1.6491	3	9531	1.4607	0.0312	110.2664	11.3189
1.6056	4	12708	1.4157	0.0625	110.4240	11.4732
1.52	5	15885	1.3731	0.125	194.1773	12.1400
1.4766	6	19062	1.3199	0.25	281.8424	12.4997
1.4006	7	22239	1.2583	0.5	460.6353	13.0648
1.3184	8.0	25416	1.1954	1.0	798.1731	13.6015
1.2527	9.0	28593	1.1609	1.0	765.5459	13.8703
1.2303	10.0	31770	1.1324	1.0	730.7526	14.1380
1.168	11.0	34947	1.1149	1.0	741.3822	14.6429
1.1215	12.0	38124	1.1003	1.0	759.3569	14.5906
1.0906	13.0	41301	1.0891	1.0	748.3090	14.9765
1.0568	14.0	44478	1.0827	1.0	750.4526	14.9267
1.0492	15.0	47655	1.0760	1.0	758.3424	14.9931
0.9851	16.0	50832	1.0698	1.0	780.3965	15.1991
0.9941	17.0	54009	1.0655	1.0	754.5295	15.1713
0.984	18.0	57186	1.0632	1.0	724.0579	15.3117
0.9459	19.0	60363	1.0621	1.0	721.8757	15.2825
0.9163	20.0	63540	1.0592	1.0	723.9400	15.3073
0.9086	21.0	66717	1.0607	1.0	725.7311	15.3085
0.8928	22.0	69894	1.0615	1.0	726.2045	15.2994
0.8819	23.0	73071	1.0590	1.0	739.2236	15.5599
0.8871	24.0	76248	1.0584	1.0	744.4550	15.3351
0.8469	25.0	79425	1.0609	1.0	724.7774	15.4515
0.8415	26.0	82602	1.0607	1.0	729.5361	15.4759
0.8151	27.0	85779	1.0664	1.0	724.0018	15.5203
0.8087	28.0	88956	1.0666	1.0	736.0040	15.4028

Framework versions

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