text_shortening_model_v67

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

• Loss: 1.2666
• Bert precision: 0.8838
• Bert recall: 0.884
• Bert f1-score: 0.8833
• Average word count: 6.5736
• Max word count: 15
• Min word count: 2
• Average token count: 10.3764
• % shortened texts with length > 12: 1.1011

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: 64
• eval_batch_size: 64
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 40

Training results

Training Loss	Epoch	Step	Validation Loss	Bert precision	Bert recall	Bert f1-score	Average word count	Max word count	Min word count	Average token count	% shortened texts with length > 12
2.9368	1.0	37	2.3418	0.7561	0.7986	0.7754	9.2703	19	0	15.1782	20.1201
2.5152	2.0	74	2.0436	0.7442	0.7935	0.767	9.2282	18	0	15.5385	16.7167
2.2465	3.0	111	1.8345	0.7787	0.8112	0.7934	8.2032	18	0	14.022	11.9119
2.0439	4.0	148	1.6984	0.8124	0.8235	0.8169	7.1001	18	0	12.2923	9.1091
1.9424	5.0	185	1.6054	0.8413	0.8423	0.8409	6.7027	18	0	11.2633	6.8068
1.8546	6.0	222	1.5450	0.8565	0.856	0.8554	6.5696	18	0	10.8759	4.7047
1.7857	7.0	259	1.5035	0.8661	0.8646	0.8646	6.5095	18	0	10.6176	3.1031
1.7348	8.0	296	1.4694	0.8718	0.8699	0.8702	6.4995	18	0	10.4224	2.6026
1.6884	9.0	333	1.4427	0.8756	0.8723	0.8733	6.4284	15	0	10.2823	2.2022
1.6823	10.0	370	1.4227	0.878	0.8753	0.8761	6.4655	15	0	10.3003	1.7017
1.6475	11.0	407	1.4069	0.8799	0.8782	0.8784	6.5375	15	1	10.3634	1.8018
1.6363	12.0	444	1.3919	0.8812	0.8797	0.8798	6.5225	15	1	10.3554	1.7017
1.6086	13.0	481	1.3784	0.8815	0.8803	0.8803	6.5105	15	1	10.3463	1.6016
1.5953	14.0	518	1.3670	0.8814	0.8802	0.8802	6.5125	15	1	10.3333	1.6016
1.5812	15.0	555	1.3569	0.8814	0.8802	0.8802	6.4955	15	1	10.3113	1.5015
1.562	16.0	592	1.3480	0.8813	0.8803	0.8802	6.5205	15	1	10.3463	1.6016
1.5541	17.0	629	1.3396	0.8817	0.8812	0.8808	6.5576	15	1	10.3764	1.4014
1.5428	18.0	666	1.3316	0.8829	0.8823	0.882	6.5495	15	1	10.3754	1.3013
1.5476	19.0	703	1.3246	0.8829	0.8821	0.8819	6.5566	15	1	10.3654	1.5015
1.5234	20.0	740	1.3169	0.8831	0.8822	0.882	6.5576	15	1	10.3794	1.4014
1.5053	21.0	777	1.3120	0.8839	0.8828	0.8827	6.5576	15	2	10.3574	1.4014
1.5	22.0	814	1.3065	0.884	0.8831	0.883	6.5606	15	2	10.3574	1.4014
1.4954	23.0	851	1.3014	0.8839	0.8833	0.883	6.5696	15	2	10.3694	1.3013
1.4875	24.0	888	1.2974	0.8838	0.8834	0.883	6.5626	15	2	10.3634	1.3013
1.4896	25.0	925	1.2941	0.8842	0.8843	0.8836	6.5826	15	2	10.3874	1.3013
1.4769	26.0	962	1.2905	0.8845	0.8844	0.8839	6.5696	15	2	10.3684	1.2012
1.4684	27.0	999	1.2864	0.8845	0.8849	0.8841	6.5886	15	2	10.3854	1.1011
1.4721	28.0	1036	1.2830	0.8843	0.8845	0.8838	6.5766	15	2	10.3654	1.1011
1.4692	29.0	1073	1.2804	0.8842	0.8844	0.8837	6.5686	15	2	10.3604	1.1011
1.4732	30.0	1110	1.2778	0.8844	0.8846	0.8839	6.5796	15	2	10.3724	1.1011
1.4592	31.0	1147	1.2754	0.8843	0.8844	0.8838	6.5646	15	2	10.3664	1.1011
1.4381	32.0	1184	1.2735	0.8844	0.8844	0.8838	6.5536	15	2	10.3524	1.1011
1.4516	33.0	1221	1.2718	0.8842	0.8842	0.8836	6.5716	15	2	10.3724	1.1011
1.4459	34.0	1258	1.2705	0.884	0.8841	0.8834	6.5746	15	2	10.3814	1.1011
1.4393	35.0	1295	1.2695	0.8838	0.8839	0.8833	6.5706	15	2	10.3784	1.1011
1.4532	36.0	1332	1.2685	0.8837	0.8839	0.8832	6.5736	15	2	10.3814	1.1011
1.4327	37.0	1369	1.2675	0.8838	0.8839	0.8833	6.5756	15	2	10.3804	1.1011
1.447	38.0	1406	1.2671	0.8838	0.884	0.8833	6.5726	15	2	10.3754	1.1011
1.4416	39.0	1443	1.2667	0.8839	0.884	0.8834	6.5756	15	2	10.3784	1.1011
1.4337	40.0	1480	1.2666	0.8838	0.884	0.8833	6.5736	15	2	10.3764	1.1011

Framework versions

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