text_shortening_model_v60

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: 0.7251
• Rouge1: 0.7246
• Rouge2: 0.5572
• Rougel: 0.6745
• Rougelsum: 0.6724
• Bert precision: 0.9227
• Bert recall: 0.9242
• Bert f1-score: 0.923
• Average word count: 8.4018
• Max word count: 16
• Min word count: 4
• Average token count: 16.1562
• % shortened texts with length > 12: 7.5893

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: 1e-05
• 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: 3

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Bert precision	Bert recall	Bert f1-score	Average word count	Max word count	Min word count	Average token count	% shortened texts with length > 12
1.4241	1.0	49	0.7533	0.7094	0.5458	0.6655	0.6641	0.9182	0.9214	0.9193	8.3884	17	5	15.3661	6.25
0.5792	2.0	98	0.7279	0.7058	0.5397	0.6587	0.6582	0.9201	0.9193	0.9192	8.3393	17	4	15.9062	5.3571
0.4392	3.0	147	0.7251	0.7246	0.5572	0.6745	0.6724	0.9227	0.9242	0.923	8.4018	16	4	16.1562	7.5893

Framework versions

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