text_shortening_model_v47

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

• Loss: 6.3912
• Rouge1: 0.0
• Rouge2: 0.0
• Rougel: 0.0
• Rougelsum: 0.0
• Bert precision: 0.6047
• Bert recall: 0.5681
• Average word count: 1.0
• Max word count: 1
• Min word count: 1
• Average token count: 12.0
• % shortened texts with length > 12: 0.0

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: 0.005
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 7

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Bert precision	Bert recall	Average word count	Max word count	Min word count	Average token count	% shortened texts with length > 12
7.822	1.0	83	7.4737	0.0776	0.0	0.0775	0.0776	0.6348	0.6223	2.0	2	2	13.0	0.0
3.2859	2.0	166	6.6585	0.1063	0.0	0.1063	0.1063	0.6469	0.608	5.0026	6	5	12.0	0.0
3.0284	3.0	249	6.4761	0.116	0.0	0.116	0.1161	0.6479	0.6388	3.9974	4	3	14.0	0.0
2.9681	4.0	332	6.4592	0.0	0.0	0.0	0.0	0.6071	0.5723	1.0	1	1	12.0	0.0
2.9377	5.0	415	6.4142	0.0	0.0	0.0	0.0	0.6047	0.5681	1.0	1	1	12.0	0.0
2.9168	6.0	498	6.4049	0.0	0.0	0.0	0.0	0.6049	0.5685	1.0	1	1	12.0	0.0
2.8964	7.0	581	6.3912	0.0	0.0	0.0	0.0	0.6047	0.5681	1.0	1	1	12.0	0.0

Framework versions

• Transformers 4.33.1
• Pytorch 2.0.1+cu118
• Datasets 2.14.5
• Tokenizers 0.13.3