text_shortening_model_v62

This model is a fine-tuned version of t5-small on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.2562
• Rouge1: 0.532
• Rouge2: 0.3351
• Rougel: 0.4834
• Rougelsum: 0.4837
• Bert precision: 0.8674
• Bert recall: 0.8593
• Bert f1-score: 0.8627
• Average word count: 8.3527
• Max word count: 16
• Min word count: 0
• Average token count: 13.2455
• % shortened texts with length > 12: 13.8393

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: 10

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
2.5656	1.0	49	2.0627	0.2906	0.1424	0.2578	0.2581	0.7433	0.7729	0.7567	9.942	18	0	16.6964	27.6786
2.1428	2.0	98	1.7604	0.2976	0.1484	0.2654	0.2647	0.7512	0.7871	0.7674	9.9062	17	1	16.6696	25.4464
1.886	3.0	147	1.5671	0.4045	0.2327	0.3717	0.3712	0.8053	0.8179	0.8102	9.2411	17	0	15.2143	25.4464
1.7092	4.0	196	1.4529	0.4575	0.2743	0.4224	0.4234	0.8309	0.8276	0.8282	8.4777	16	0	14.1339	17.4107
1.61	5.0	245	1.3795	0.4869	0.2867	0.4435	0.4445	0.8476	0.8402	0.8431	8.4196	17	0	13.5982	17.4107
1.5541	6.0	294	1.3272	0.5085	0.3053	0.466	0.4664	0.857	0.8473	0.8514	8.2455	17	0	13.2768	16.5179
1.5157	7.0	343	1.2940	0.5227	0.3225	0.4752	0.4763	0.8583	0.8504	0.8537	8.2946	17	0	13.2679	14.7321
1.456	8.0	392	1.2721	0.5272	0.3269	0.4782	0.479	0.8653	0.857	0.8605	8.3839	17	0	13.2411	14.2857
1.4422	9.0	441	1.2600	0.527	0.3315	0.4793	0.4807	0.8656	0.8576	0.8609	8.3304	16	0	13.2679	13.8393
1.4384	10.0	490	1.2562	0.532	0.3351	0.4834	0.4837	0.8674	0.8593	0.8627	8.3527	16	0	13.2455	13.8393

Framework versions

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