08cac0e7fc2540de88af08514b4cd03e

This model is a fine-tuned version of studio-ousia/mluke-base on the fancyzhx/dbpedia_14 dataset. It achieves the following results on the evaluation set:

• Loss: 0.0577
• Data Size: 1.0
• Epoch Runtime: 1586.1733
• Accuracy: 0.9893
• F1 Macro: 0.9893
• Rouge1: 0.9893
• Rouge2: 0.0
• Rougel: 0.9893
• Rougelsum: 0.9893

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	Accuracy	F1 Macro	Rouge1	Rouge2	Rougel	Rougelsum
No log	0	0	2.6273	0	54.8888	0.0830	0.0440	0.0831	0.0	0.0830	0.0831
0.1405	1	17500	0.0760	0.0078	67.3486	0.9838	0.9839	0.9838	0.0	0.9838	0.9838
0.0611	2	35000	0.0727	0.0156	79.5771	0.9857	0.9857	0.9857	0.0	0.9857	0.9857
0.0641	3	52500	0.0825	0.0312	103.3333	0.9836	0.9835	0.9836	0.0	0.9836	0.9836
0.0614	4	70000	0.0638	0.0625	150.7420	0.9867	0.9867	0.9867	0.0	0.9866	0.9867
0.0472	5	87500	0.0638	0.125	244.0766	0.9871	0.9870	0.9871	0.0	0.9870	0.9871
0.0591	6	105000	0.0506	0.25	430.5615	0.9887	0.9887	0.9887	0.0	0.9887	0.9887
0.0004	7	122500	0.0464	0.5	817.6506	0.9897	0.9897	0.9898	0.0	0.9897	0.9897
0.042	8.0	140000	0.0485	1.0	1586.6349	0.9906	0.9906	0.9907	0.0	0.9907	0.9906
0.0325	9.0	157500	0.0513	1.0	1582.4599	0.9897	0.9897	0.9897	0.0	0.9897	0.9897
0.0216	10.0	175000	0.0649	1.0	1581.0572	0.9890	0.9889	0.9890	0.0	0.9889	0.9890
0.0431	11.0	192500	0.0577	1.0	1586.1733	0.9893	0.9893	0.9893	0.0	0.9893	0.9893

Framework versions

• Transformers 4.57.0
• Pytorch 2.8.0+cu128
• Datasets 4.3.0
• Tokenizers 0.22.1