PhysicalScienceBARTMainSections

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

• Loss: 4.2611
• Rouge1: 53.3257
• Rouge2: 19.9372
• Rougel: 38.7516
• Rougelsum: 49.5491
• Bertscore Precision: 82.9683
• Bertscore Recall: 84.3765
• Bertscore F1: 83.6629
• Bleu: 0.1444
• Gen Len: 195.4093

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: 16
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 1

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Bertscore Precision	Bertscore Recall	Bertscore F1	Bleu	Gen Len
6.0567	0.0622	100	5.9024	44.7542	14.8626	32.3438	41.6627	79.6577	81.9972	80.8049	0.1058	195.4093
5.628	0.1244	200	5.5009	44.7374	15.4406	32.7203	41.5684	79.7952	82.4154	81.0775	0.1106	195.4093
5.3608	0.1866	300	5.2016	47.9813	16.6932	34.1908	44.4923	80.6116	82.8487	81.709	0.1189	195.4093
5.1172	0.2489	400	5.0247	49.6117	17.0694	35.1947	46.0939	81.4181	83.2142	82.3018	0.1228	195.4093
5.1058	0.3111	500	4.8769	49.7791	17.282	35.3202	45.4459	80.9748	83.2981	82.1135	0.1250	195.4093
4.9831	0.3733	600	4.7486	49.7885	17.5964	36.1885	46.1291	81.8182	83.5683	82.6792	0.1263	195.4093
4.7239	0.4355	700	4.6365	49.9977	18.0061	36.4943	46.3477	81.7979	83.6503	82.7089	0.1299	195.4093
4.6893	0.4977	800	4.5773	51.7141	18.7056	37.2897	48.1051	82.4355	83.9204	83.1676	0.1347	195.4093
4.641	0.5599	900	4.5179	51.337	18.6106	37.3188	47.6183	82.1666	83.9203	83.0297	0.1355	195.4093
4.4518	0.6222	1000	4.4457	52.5898	18.9865	37.7363	48.9758	82.5619	84.028	83.2849	0.1363	195.4093
4.4246	0.6844	1100	4.4001	52.5771	19.1928	37.92	48.9098	82.5426	84.0673	83.2942	0.1392	195.4093
4.549	0.7466	1200	4.3539	52.3117	19.2452	38.0721	48.7304	82.7547	84.1096	83.4232	0.1383	195.4093
4.3528	0.8088	1300	4.3296	52.5899	19.6953	38.3709	48.8248	82.7757	84.2642	83.5094	0.1424	195.4093
4.3692	0.8710	1400	4.2972	53.2821	19.763	38.4702	49.3072	82.9332	84.4115	83.6622	0.1434	195.4093
4.2056	0.9332	1500	4.2795	53.4962	19.9871	38.7098	49.5905	83.0029	84.4307	83.7072	0.1449	195.4093
4.3956	0.9955	1600	4.2611	53.3257	19.9372	38.7516	49.5491	82.9683	84.3765	83.6629	0.1444	195.4093

Framework versions

• Transformers 4.41.2
• Pytorch 2.3.1+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1