SocialScienceBART

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: 4.6504
• Rouge1: 51.3376
• Rouge2: 18.2656
• Rougel: 36.0279
• Rougelsum: 47.688
• Bertscore Precision: 81.2268
• Bertscore Recall: 83.5394
• Bertscore F1: 82.3632
• Bleu: 0.1466
• Gen Len: 195.1837

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
6.3499	0.1314	100	5.9601	44.2626	14.89	31.4939	41.4297	78.5226	81.7375	80.0922	0.1200	195.1837
5.7616	0.2628	200	5.5234	45.0397	15.5203	31.9711	41.5206	77.988	82.1682	80.0147	0.1261	195.1837
5.3197	0.3943	300	5.2264	46.0652	15.9668	32.867	42.5272	78.4756	82.4394	80.4011	0.1308	195.1837
5.1661	0.5257	400	5.0219	45.5622	15.8452	33.3135	42.7801	79.6663	82.5824	81.0931	0.1287	195.1837
5.0513	0.6571	500	4.8896	45.2597	15.7552	33.7344	42.337	79.3705	82.7284	81.0087	0.1287	195.1837
4.8073	0.7885	600	4.7506	48.6142	17.418	35.1837	45.2098	80.4041	83.2297	81.7876	0.1409	195.1837
4.7293	0.9199	700	4.6504	51.3376	18.2656	36.0279	47.688	81.2268	83.5394	82.3632	0.1466	195.1837

Framework versions

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