led-risalah_data_v2

This model was trained from scratch on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.6785
• Rouge1 Precision: 0.6665
• Rouge1 Recall: 0.1816
• Rouge1 Fmeasure: 0.284

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: 1
• eval_batch_size: 1
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 8
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1 Fmeasure	Rouge1 Precision	Rouge1 Recall
2.4597	0.91	8	1.8034	0.1951	0.4699	0.1246
1.7706	1.94	17	1.6403	0.2451	0.6043	0.1554
1.5072	2.97	26	1.5947	0.2628	0.6236	0.1676
1.4018	4.0	35	1.5688	0.2789	0.656	0.1782
1.2761	4.91	43	1.5454	0.2723	0.6434	0.1736
1.1779	5.94	52	1.5636	0.2889	0.6794	0.1843
1.1235	6.97	61	1.5430	0.2965	0.6913	0.1902
1.0529	8.0	70	1.5639	0.2829	0.6705	0.1805
0.9883	8.91	78	1.5740	0.2817	0.6757	0.1798
0.9274	9.94	87	1.5793	0.2771	0.6623	0.1764
0.925	10.97	96	1.6072	0.2821	0.665	0.18
0.858	12.0	105	1.6129	0.284	0.6625	0.1817
0.8182	12.91	113	1.6396	0.2765	0.6567	0.1761
0.7974	13.94	122	1.6445	0.2759	0.659	0.1759
0.7524	14.97	131	1.6585	0.2763	0.6585	0.1759
0.7743	16.0	140	1.6779	0.2788	0.6594	0.1779
0.7486	16.91	148	1.6742	0.2851	0.6666	0.1819
0.676	17.94	157	1.6790	0.2859	0.6707	0.1827

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.1+cu121
• Datasets 2.14.5
• Tokenizers 0.15.1