BART-Large-psychological-dataset

This model is a fine-tuned version of facebook/bart-large-cnn on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.1549
• Rouge1: 0.6621
• Rouge2: 0.4488
• Rougel: 0.5658
• Rougelsum: 0.5656
• Gen Len: 80.6204

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: 2e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 8
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
No log	1.0	274	0.8651	0.6252	0.3953	0.5206	0.5206	86.9982
0.7484	2.0	548	0.8332	0.648	0.4301	0.554	0.5541	79.885
0.7484	3.0	822	0.8943	0.6498	0.4335	0.5514	0.5518	82.635
0.3207	4.0	1096	0.9653	0.6571	0.4422	0.5607	0.5609	79.9708
0.3207	5.0	1370	1.0514	0.6582	0.4445	0.5637	0.5639	79.8047
0.1557	6.0	1644	1.0752	0.6607	0.4476	0.5659	0.5657	79.6058
0.1557	7.0	1918	1.1302	0.6588	0.4443	0.5626	0.5626	80.5821
0.0845	8.0	2192	1.1549	0.6621	0.4488	0.5658	0.5656	80.6204

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.1+cu121
• Datasets 3.0.0
• Tokenizers 0.19.1