d5ee92af64f5995dfb2cff04cdd13fc7

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: 1.4281
• Data Size: 1.0
• Epoch Runtime: 198.4646
• Bleu: 9.9183

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	3.3889	0	19.5402	0.5448
No log	1	808	3.2189	0.0078	23.3609	0.7044
No log	2	1616	3.0458	0.0156	29.5394	1.1133
No log	3	2424	2.8911	0.0312	25.6167	1.7552
0.0977	4	3232	2.7318	0.0625	32.6952	2.7459
2.9899	5	4040	2.5805	0.125	42.8306	3.3725
2.7465	6	4848	2.4096	0.25	78.0312	3.9332
2.4591	7	5656	2.2130	0.5	108.7962	4.8950
2.2598	8.0	6464	2.0027	1.0	204.9841	5.9369
2.0827	9.0	7272	1.8715	1.0	201.9693	6.6774
1.9699	10.0	8080	1.7852	1.0	202.4802	7.3452
1.8979	11.0	8888	1.7264	1.0	200.1101	7.4456
1.806	12.0	9696	1.6791	1.0	197.1412	7.8435
1.731	13.0	10504	1.6473	1.0	190.5002	8.0464
1.686	14.0	11312	1.6093	1.0	188.9600	8.2376
1.6499	15.0	12120	1.5840	1.0	188.9396	8.4719
1.636	16.0	12928	1.5622	1.0	189.0705	8.6122
1.5533	17.0	13736	1.5403	1.0	183.1241	8.7637
1.5276	18.0	14544	1.5243	1.0	183.3175	8.8833
1.4875	19.0	15352	1.5121	1.0	207.7781	9.1169
1.4468	20.0	16160	1.4963	1.0	200.3920	9.1539
1.4129	21.0	16968	1.4895	1.0	187.3135	9.3170
1.3828	22.0	17776	1.4767	1.0	194.1532	9.4011
1.357	23.0	18584	1.4679	1.0	193.1984	9.4004
1.3232	24.0	19392	1.4631	1.0	206.8371	9.5002
1.303	25.0	20200	1.4546	1.0	189.9689	9.5300
1.2748	26.0	21008	1.4482	1.0	205.0709	9.7769
1.2569	27.0	21816	1.4486	1.0	200.3493	9.6726
1.2345	28.0	22624	1.4458	1.0	189.3574	9.6684
1.2344	29.0	23432	1.4407	1.0	192.3539	9.7323
1.1679	30.0	24240	1.4383	1.0	210.2699	9.7481
1.1749	31.0	25048	1.4340	1.0	192.7680	9.6688
1.1594	32.0	25856	1.4361	1.0	221.8852	9.7711
1.1395	33.0	26664	1.4381	1.0	192.2963	9.7037
1.1177	34.0	27472	1.4270	1.0	190.7512	9.7191
1.0783	35.0	28280	1.4329	1.0	189.8960	9.7003
1.1051	36.0	29088	1.4316	1.0	216.8254	9.6882
1.0664	37.0	29896	1.4326	1.0	194.1363	9.7428
1.056	38.0	30704	1.4281	1.0	198.4646	9.9183

Framework versions

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