text_shortening_model_v4

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

• Loss: 1.4263
• Rouge1: 0.587
• Rouge2: 0.3563
• Rougel: 0.5367
• Rougelsum: 0.5356
• Bert precision: 0.8882
• Bert recall: 0.9005
• Average word count: 11.8286
• Max word count: 18
• Min word count: 6
• Average token count: 17.0929

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: 20

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
1.3135	1.0	8	1.8236	0.5468	0.3281	0.4997	0.4987	0.8808	0.886	11.5786	18	6	16.8286
1.1741	2.0	16	1.6858	0.5482	0.3138	0.4936	0.4934	0.8776	0.8889	12.1429	18	5	17.2929
1.1284	3.0	24	1.6250	0.5601	0.3292	0.5053	0.5053	0.8817	0.8922	12.0357	18	5	17.0786
1.1142	4.0	32	1.5850	0.5645	0.3397	0.5164	0.516	0.8839	0.8954	11.9357	18	4	17.0571
1.0745	5.0	40	1.5500	0.5777	0.3465	0.5272	0.5263	0.8863	0.8995	12.1071	18	4	17.2143
1.0354	6.0	48	1.5235	0.5796	0.3451	0.5263	0.5252	0.8859	0.8992	12.0	18	5	17.1
1.0126	7.0	56	1.5026	0.5859	0.3509	0.53	0.5291	0.8873	0.8998	11.8786	18	5	17.0714
1.0087	8.0	64	1.4877	0.5828	0.3511	0.5323	0.5304	0.8869	0.8989	11.8143	18	6	16.9857
0.9745	9.0	72	1.4758	0.5879	0.3533	0.5343	0.5332	0.8874	0.9008	11.8857	18	6	17.0786
0.9712	10.0	80	1.4638	0.585	0.3532	0.5319	0.5303	0.8878	0.9007	11.8643	18	6	17.0643
0.9556	11.0	88	1.4567	0.5909	0.3546	0.5348	0.5336	0.8879	0.9014	11.9357	18	6	17.1571
0.9413	12.0	96	1.4540	0.5881	0.3533	0.5351	0.5342	0.8879	0.9015	11.9571	18	6	17.25
0.9344	13.0	104	1.4489	0.5904	0.3602	0.5388	0.5374	0.8879	0.9013	11.9714	18	6	17.2643
0.929	14.0	112	1.4399	0.5866	0.355	0.5348	0.5338	0.8877	0.9006	11.8929	18	6	17.1857
0.9118	15.0	120	1.4353	0.5885	0.3569	0.537	0.5362	0.8883	0.9004	11.8	18	6	17.0857
0.9075	16.0	128	1.4326	0.5862	0.3531	0.5337	0.5329	0.8875	0.8998	11.8286	18	6	17.1143
0.9217	17.0	136	1.4296	0.5841	0.3547	0.534	0.5331	0.8882	0.9	11.7929	18	6	17.0571
0.8936	18.0	144	1.4270	0.5856	0.3558	0.5356	0.5347	0.8888	0.9003	11.75	18	6	17.0143
0.8848	19.0	152	1.4262	0.587	0.3564	0.5369	0.5357	0.8884	0.9005	11.8214	18	6	17.0857
0.8913	20.0	160	1.4263	0.587	0.3563	0.5367	0.5356	0.8882	0.9005	11.8286	18	6	17.0929

Framework versions

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