e7f4a25c2e8a2a010f1b5664c9543133

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

• Loss: 1.6011
• Data Size: 1.0
• Epoch Runtime: 95.5761
• Bleu: 6.9557

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	3.5882	0	15.7346	0.7908
No log	1	390	3.3958	0.0078	10.1846	0.8633
No log	2	780	3.2440	0.0156	10.5299	1.0879
No log	3	1170	3.1277	0.0312	15.6032	1.2929
No log	4	1560	3.0035	0.0625	14.8827	1.3321
0.1859	5	1950	2.8599	0.125	26.4357	1.3222
0.4554	6	2340	2.6978	0.25	34.8773	2.0002
2.7876	7	2730	2.5020	0.5	52.5474	2.8249
2.5462	8.0	3120	2.2868	1.0	92.1808	3.4739
2.3654	9.0	3510	2.1553	1.0	91.9900	3.9402
2.2479	10.0	3900	2.0582	1.0	116.6953	4.3567
2.1552	11.0	4290	1.9875	1.0	102.0627	4.6126
2.0659	12.0	4680	1.9283	1.0	103.3583	4.8501
2.0287	13.0	5070	1.8807	1.0	93.8514	5.0581
1.9624	14.0	5460	1.8412	1.0	99.0481	5.1799
1.8765	15.0	5850	1.8109	1.0	94.2898	5.4336
1.8305	16.0	6240	1.7812	1.0	101.9515	5.6854
1.7974	17.0	6630	1.7529	1.0	96.1651	5.6819
1.7448	18.0	7020	1.7374	1.0	95.5570	5.8152
1.7238	19.0	7410	1.7175	1.0	103.9164	5.9628
1.6954	20.0	7800	1.7036	1.0	93.7302	5.9723
1.6776	21.0	8190	1.6894	1.0	104.0247	6.1403
1.6067	22.0	8580	1.6673	1.0	96.3601	6.1746
1.5901	23.0	8970	1.6637	1.0	103.8547	6.2606
1.561	24.0	9360	1.6508	1.0	95.8741	6.3408
1.5278	25.0	9750	1.6448	1.0	97.7640	6.4218
1.507	26.0	10140	1.6358	1.0	107.6500	6.5314
1.4703	27.0	10530	1.6249	1.0	95.0464	6.5072
1.4578	28.0	10920	1.6216	1.0	92.9810	6.5518
1.4404	29.0	11310	1.6139	1.0	93.4152	6.6541
1.4162	30.0	11700	1.6078	1.0	91.5107	6.5969
1.3787	31.0	12090	1.6158	1.0	97.0743	6.6742
1.3495	32.0	12480	1.6000	1.0	90.7719	6.7545
1.3391	33.0	12870	1.6004	1.0	92.2922	6.8086
1.3143	34.0	13260	1.5971	1.0	89.4478	6.7237
1.296	35.0	13650	1.5995	1.0	92.8490	6.7595
1.2918	36.0	14040	1.5946	1.0	94.9193	6.7271
1.26	37.0	14430	1.5945	1.0	92.7881	6.8017
1.2362	38.0	14820	1.5960	1.0	92.6619	6.8436
1.2117	39.0	15210	1.5950	1.0	91.1050	6.8313
1.1918	40.0	15600	1.5939	1.0	92.5751	6.8792
1.1852	41.0	15990	1.5950	1.0	94.1525	6.9230
1.1906	42.0	16380	1.5960	1.0	95.4994	6.8665
1.1515	43.0	16770	1.6002	1.0	96.5721	6.9423
1.1448	44.0	17160	1.6011	1.0	95.5761	6.9557

Framework versions

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