decafd18d2dee1b3cd3d3a28a6635305

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: 0.9814
• Data Size: 1.0
• Epoch Runtime: 98.1355
• Bleu: 13.3563

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	2.5607	0	14.5991	0.1842
No log	1	437	2.4747	0.0078	16.6316	0.2266
No log	2	874	2.3841	0.0156	12.4275	0.4988
No log	3	1311	2.2488	0.0312	17.1922	0.5118
No log	4	1748	2.1029	0.0625	19.2542	0.7667
2.2109	5	2185	1.9879	0.125	26.4880	1.2838
2.1006	6	2622	1.8642	0.25	42.0241	2.6669
1.9291	7	3059	1.7206	0.5	70.2491	4.0708
1.7381	8.0	3496	1.5636	1.0	149.8675	5.5590
1.6181	9.0	3933	1.4499	1.0	130.7932	6.6572
1.5056	10.0	4370	1.3723	1.0	145.4794	7.4242
1.447	11.0	4807	1.3099	1.0	145.7462	8.0778
1.383	12.0	5244	1.2594	1.0	108.3039	8.8326
1.3299	13.0	5681	1.2202	1.0	115.9984	9.1728
1.2613	14.0	6118	1.1907	1.0	104.9783	9.7938
1.2273	15.0	6555	1.1621	1.0	102.6652	9.7652
1.1678	16.0	6992	1.1362	1.0	100.6030	10.2138
1.1441	17.0	7429	1.1101	1.0	105.9486	10.3005
1.1065	18.0	7866	1.0993	1.0	106.0890	11.0051
1.0658	19.0	8303	1.0737	1.0	104.2994	10.9763
1.0488	20.0	8740	1.0597	1.0	101.6413	11.4335
1.0181	21.0	9177	1.0459	1.0	101.3497	11.7183
0.9851	22.0	9614	1.0368	1.0	101.2914	11.7005
0.963	23.0	10051	1.0302	1.0	111.7042	11.8514
0.9359	24.0	10488	1.0192	1.0	111.2087	12.1365
0.9233	25.0	10925	1.0095	1.0	105.1604	12.3197
0.9022	26.0	11362	1.0089	1.0	104.0405	12.3111
0.8816	27.0	11799	1.0057	1.0	104.8092	12.5789
0.8525	28.0	12236	0.9977	1.0	106.6144	12.4490
0.8293	29.0	12673	0.9884	1.0	98.7457	12.7247
0.8088	30.0	13110	0.9838	1.0	100.4866	12.8502
0.8072	31.0	13547	0.9825	1.0	98.6900	12.8563
0.7987	32.0	13984	0.9789	1.0	99.3213	12.8370
0.7723	33.0	14421	0.9761	1.0	98.1782	12.9031
0.7595	34.0	14858	0.9773	1.0	97.5901	13.0862
0.7414	35.0	15295	0.9782	1.0	102.2905	13.1639
0.7138	36.0	15732	0.9729	1.0	96.9078	13.5037
0.7123	37.0	16169	0.9727	1.0	97.9281	13.1368
0.6947	38.0	16606	0.9816	1.0	101.6442	13.5031
0.6734	39.0	17043	0.9750	1.0	99.8666	13.1697
0.664	40.0	17480	0.9799	1.0	98.9188	13.2457
0.6697	41.0	17917	0.9814	1.0	98.1355	13.3563

Framework versions

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