results_pegasus2-_wiki

This model is a fine-tuned version of google/pegasus-large on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0771
• Rouge1: 0.2644
• Rouge2: 0.1159
• Rougel: 0.264
• Rougelsum: 0.2635
• Gen Len: 248.7564

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: 4e-05
• train_batch_size: 2
• eval_batch_size: 2
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 4
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 250
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
2.6037	0.5222	500	0.2910	0.0	0.0	0.0	0.0	223.1886
0.2787	1.0444	1000	0.2403	0.0515	0.0	0.0525	0.052	221.6974
0.2284	1.5666	1500	0.1922	0.0607	0.0	0.0621	0.0614	246.9666
0.1827	2.0888	2000	0.1775	0.1271	0.0176	0.129	0.1278	247.4322
0.167	2.6110	2500	0.1437	0.1591	0.0347	0.1597	0.1598	248.3084
0.1537	3.1332	3000	0.1301	0.1765	0.047	0.1765	0.1754	249.0864
0.14	3.6554	3500	0.1183	0.2082	0.059	0.2086	0.2077	248.2633
0.1306	4.1775	4000	0.1092	0.2095	0.0599	0.209	0.2083	246.5972
0.1272	4.6997	4500	0.1024	0.2181	0.0719	0.2177	0.2172	247.3752
0.1177	5.2219	5000	0.1013	0.2217	0.0725	0.2217	0.2211	247.4224
0.1123	5.7441	5500	0.0929	0.2242	0.0797	0.2249	0.2243	247.277
0.1114	6.2663	6000	0.0887	0.2335	0.0839	0.2334	0.233	247.3399
0.1073	6.7885	6500	0.0835	0.2452	0.0976	0.2461	0.2452	249.2043
0.1025	7.3107	7000	0.0821	0.2458	0.0971	0.2456	0.2455	246.2063
0.1009	7.8329	7500	0.0821	0.251	0.1009	0.2509	0.2508	248.7642
0.1004	8.3551	8000	0.0834	0.2583	0.1058	0.2587	0.258	248.7525
0.0965	8.8773	8500	0.0791	0.2621	0.116	0.2622	0.2621	248.7407
0.0975	9.3995	9000	0.0781	0.2619	0.1147	0.2613	0.2608	248.4185
0.0941	9.9217	9500	0.0771	0.2644	0.1159	0.264	0.2635	248.7564

Framework versions

• Transformers 4.42.0.dev0
• Pytorch 2.3.0+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1