PhysicalScienceBARTPrincipal

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: 4.5862
• Rouge1: 49.7214
• Rouge2: 15.9205
• Rougel: 34.8099
• Rougelsum: 45.9442
• Bertscore Precision: 81.8626
• Bertscore Recall: 83.3072
• Bertscore F1: 82.5744
• Bleu: 0.1065
• Gen Len: 196.3779

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
6.4881	0.0620	100	6.2790	38.9402	10.9737	28.0473	36.4124	78.6712	80.7927	79.7123	0.0702	196.3779
5.9838	0.1239	200	5.8574	39.6094	11.61	28.6653	36.6426	78.5563	81.2374	79.8672	0.0773	196.3779
5.5757	0.1859	300	5.5425	43.235	12.5595	30.3069	40.1431	79.7016	81.7103	80.6878	0.0826	196.3779
5.4752	0.2478	400	5.3518	45.0647	13.1878	31.0925	41.4826	79.7122	82.0455	80.8554	0.0880	196.3779
5.3711	0.3098	500	5.2193	47.1793	13.5223	31.7989	43.5774	80.6424	82.3476	81.4813	0.0892	196.3779
5.1653	0.3717	600	5.0858	45.2081	13.4909	31.8919	41.7813	80.7104	82.4561	81.5689	0.0897	196.3779
5.0684	0.4337	700	4.9837	46.4035	14.2034	32.654	42.8883	80.4628	82.4529	81.4399	0.0941	196.3779
4.9625	0.4957	800	4.9084	48.2088	14.8904	33.2025	44.5397	81.1668	82.8469	81.9935	0.0986	196.3779
4.8858	0.5576	900	4.8370	48.5919	14.7721	33.5041	44.7923	81.2656	82.8635	82.0522	0.0974	196.3779
4.8251	0.6196	1000	4.7813	49.2512	15.4584	34.0164	45.5215	81.4958	83.0067	82.2398	0.1030	196.3779
4.8581	0.6815	1100	4.7307	48.7203	15.379	34.0451	45.0395	81.7154	83.106	82.4008	0.1027	196.3779
4.7934	0.7435	1200	4.6861	49.5987	15.6207	34.3261	45.8512	81.7656	83.1546	82.4502	0.1042	196.3779
4.7163	0.8055	1300	4.6518	48.9818	15.5333	34.3788	45.3444	81.6763	83.1451	82.3998	0.1039	196.3779
4.6855	0.8674	1400	4.6199	49.1462	15.5914	34.5149	45.5788	81.7027	83.1199	82.401	0.1037	196.3779
4.615	0.9294	1500	4.5987	49.6903	15.8973	34.7628	45.9111	81.8545	83.302	82.5678	0.1064	196.3779
4.5964	0.9913	1600	4.5862	49.7214	15.9205	34.8099	45.9442	81.8626	83.3072	82.5744	0.1065	196.3779

Framework versions

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