text_shortening_model_v5

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

• Loss: 1.3950
• Rouge1: 0.6032
• Rouge2: 0.3745
• Rougel: 0.5559
• Rougelsum: 0.556
• Bert precision: 0.8961
• Bert recall: 0.9059
• Average word count: 11.4071
• Max word count: 16
• Min word count: 6
• Average token count: 16.7643

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: 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: 50

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
1.311	1.0	8	1.8181	0.5439	0.3249	0.4963	0.4961	0.879	0.8847	11.65	18	1	16.8857
1.174	2.0	16	1.6800	0.55	0.3147	0.4935	0.4931	0.8779	0.8891	12.1214	18	5	17.2857
1.1265	3.0	24	1.6149	0.5642	0.3349	0.5109	0.5105	0.8833	0.8935	11.8643	18	5	16.9571
1.1075	4.0	32	1.5730	0.5657	0.3383	0.5163	0.5161	0.8836	0.8961	11.9643	18	4	17.0929
1.062	5.0	40	1.5421	0.5819	0.3544	0.53	0.5292	0.8858	0.9007	12.1286	18	5	17.2571
1.021	6.0	48	1.5085	0.5792	0.3514	0.5262	0.5255	0.8848	0.8986	11.9929	18	5	17.1
0.998	7.0	56	1.4826	0.5825	0.3548	0.5335	0.5317	0.887	0.9	11.8357	18	6	17.0857
0.9794	8.0	64	1.4659	0.5814	0.3508	0.5306	0.5297	0.8877	0.8993	11.6714	18	4	16.9286
0.9553	9.0	72	1.4533	0.5871	0.3545	0.533	0.5318	0.8874	0.9018	11.8857	18	6	17.2071
0.9451	10.0	80	1.4402	0.5871	0.3604	0.5368	0.5361	0.8889	0.9013	11.6571	18	6	16.9929
0.9223	11.0	88	1.4334	0.5888	0.3602	0.5378	0.5369	0.8883	0.9017	11.8071	18	6	17.1643
0.893	12.0	96	1.4295	0.587	0.3589	0.5367	0.5356	0.8878	0.9008	11.8	18	6	17.1214
0.8768	13.0	104	1.4182	0.5887	0.3598	0.5395	0.5388	0.8887	0.9021	11.8571	17	6	17.2429
0.8598	14.0	112	1.4076	0.5937	0.3647	0.5476	0.5466	0.8909	0.9021	11.6214	16	6	16.9429
0.8555	15.0	120	1.4080	0.5948	0.3668	0.5481	0.5473	0.89	0.9018	11.6786	16	6	17.0429
0.8505	16.0	128	1.4067	0.5984	0.3705	0.5517	0.5507	0.8908	0.9031	11.7214	17	6	17.0714
0.8545	17.0	136	1.3995	0.5946	0.3669	0.5479	0.547	0.8924	0.9028	11.55	15	6	16.9071
0.8025	18.0	144	1.3953	0.5935	0.3637	0.547	0.5461	0.8924	0.9022	11.5571	15	6	16.8929
0.7915	19.0	152	1.3975	0.5963	0.3702	0.5485	0.5476	0.8899	0.9025	11.7714	17	6	17.1929
0.8017	20.0	160	1.3957	0.5915	0.3633	0.5439	0.542	0.8897	0.902	11.7143	17	6	17.1643
0.8133	21.0	168	1.3926	0.5932	0.3632	0.5438	0.5425	0.8916	0.9022	11.5714	16	6	16.9786
0.7858	22.0	176	1.3942	0.5941	0.3658	0.5453	0.544	0.8915	0.9022	11.5714	16	6	16.9857
0.7712	23.0	184	1.3929	0.6015	0.3698	0.5506	0.5498	0.8916	0.9044	11.7714	16	6	17.1786
0.7786	24.0	192	1.3900	0.5985	0.3662	0.549	0.5482	0.8926	0.903	11.5286	16	6	16.8857
0.7707	25.0	200	1.3888	0.6011	0.3708	0.5508	0.5495	0.8947	0.9037	11.3786	15	6	16.7286
0.7661	26.0	208	1.3888	0.6001	0.3704	0.5512	0.55	0.8943	0.9033	11.4429	15	6	16.8
0.7489	27.0	216	1.3892	0.5953	0.3673	0.5467	0.5462	0.8927	0.9017	11.4429	15	6	16.7929
0.7433	28.0	224	1.3910	0.5925	0.3661	0.5449	0.5449	0.8927	0.9023	11.4714	15	6	16.9
0.7295	29.0	232	1.3886	0.5934	0.3656	0.5458	0.5451	0.893	0.9019	11.4929	15	6	16.8429
0.7446	30.0	240	1.3874	0.5947	0.3643	0.5474	0.5471	0.893	0.9017	11.4929	15	6	16.7786
0.7318	31.0	248	1.3848	0.5998	0.3708	0.5518	0.5517	0.8946	0.9029	11.5	15	6	16.7714
0.7279	32.0	256	1.3851	0.6003	0.3703	0.5522	0.5522	0.8948	0.9035	11.5214	15	6	16.7929
0.725	33.0	264	1.3879	0.5979	0.3677	0.5487	0.5476	0.8956	0.9046	11.4643	15	6	16.7214
0.7229	34.0	272	1.3907	0.5959	0.3677	0.5463	0.5457	0.8948	0.904	11.5286	15	6	16.8143
0.7228	35.0	280	1.3916	0.5983	0.3696	0.5499	0.5491	0.8947	0.9047	11.5857	15	6	16.8714
0.7006	36.0	288	1.3913	0.5962	0.3681	0.5461	0.5454	0.8938	0.9036	11.5571	15	6	16.8286
0.6935	37.0	296	1.3891	0.5976	0.3707	0.55	0.5496	0.895	0.9042	11.3786	15	6	16.6857
0.7011	38.0	304	1.3894	0.602	0.3727	0.5546	0.554	0.8965	0.9059	11.4429	16	6	16.6929
0.7188	39.0	312	1.3903	0.6031	0.373	0.5556	0.5548	0.896	0.9061	11.5357	16	6	16.7929
0.7013	40.0	320	1.3927	0.6055	0.3763	0.5573	0.5564	0.8952	0.906	11.5929	16	6	16.8929
0.6857	41.0	328	1.3932	0.5991	0.3729	0.5509	0.5514	0.894	0.9054	11.5357	16	6	16.8857
0.7063	42.0	336	1.3933	0.5995	0.3739	0.5514	0.5513	0.8943	0.9056	11.5571	16	6	16.8571
0.7022	43.0	344	1.3935	0.5974	0.3714	0.55	0.5503	0.894	0.9052	11.55	16	6	16.8714
0.6975	44.0	352	1.3937	0.6008	0.369	0.5519	0.5516	0.8949	0.905	11.5286	16	6	16.8071
0.687	45.0	360	1.3937	0.6024	0.3705	0.5536	0.5534	0.8955	0.9053	11.4929	16	6	16.7786
0.7044	46.0	368	1.3944	0.6024	0.3718	0.5545	0.5543	0.8957	0.9054	11.4643	16	6	16.7714
0.695	47.0	376	1.3947	0.6037	0.3746	0.5558	0.5556	0.896	0.9059	11.45	16	6	16.7857
0.7019	48.0	384	1.3949	0.6047	0.3756	0.5575	0.5572	0.896	0.9058	11.4357	16	6	16.7643
0.6895	49.0	392	1.3950	0.6032	0.3745	0.5559	0.556	0.8961	0.9059	11.4071	16	6	16.7643
0.6914	50.0	400	1.3950	0.6032	0.3745	0.5559	0.556	0.8961	0.9059	11.4071	16	6	16.7643

Framework versions

• Transformers 4.32.1
• Pytorch 2.0.1+cu118
• Datasets 2.14.4
• Tokenizers 0.13.3