text_shortening_model_v49

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: 1.7760
• Rouge1: 0.5119
• Rouge2: 0.2768
• Rougel: 0.4448
• Rougelsum: 0.4444
• Bert precision: 0.8755
• Bert recall: 0.8801
• Average word count: 8.8492
• Max word count: 20
• Min word count: 5
• Average token count: 16.4709
• % shortened texts with length > 12: 8.7302

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: 0.0001
• 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: 5

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
1.8542	1.0	83	1.6189	0.5121	0.2699	0.4302	0.4304	0.863	0.8909	11.3386	21	5	19.4312	31.746
0.9651	2.0	166	1.4837	0.4957	0.2664	0.4347	0.4362	0.8687	0.8758	8.8598	19	4	16.9815	9.2593
0.608	3.0	249	1.4074	0.5012	0.2693	0.4346	0.4342	0.8725	0.8781	8.836	20	4	15.5265	5.5556
0.3788	4.0	332	1.5646	0.5202	0.2836	0.4535	0.4537	0.876	0.881	8.9312	18	5	16.4365	10.3175
0.2296	5.0	415	1.7760	0.5119	0.2768	0.4448	0.4444	0.8755	0.8801	8.8492	20	5	16.4709	8.7302

Framework versions

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