t5-small-wsd-finetuned-cve-reason

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

• Loss: 0.2766
• Rouge1: 89.4973
• Rouge2: 85.3832
• Rougel: 89.485
• Rougelsum: 89.3862
• Gen Len: 7.3574

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: 2e-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: 200
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
No log	1.0	56	2.7012	21.9778	11.8706	21.6107	21.6246	18.5607
No log	2.0	112	1.5470	42.288	30.0431	42.0451	41.9627	11.0426
No log	3.0	168	1.1416	63.7927	48.1638	63.8356	63.7446	7.2557
No log	4.0	224	0.9195	67.0177	53.3701	66.9943	66.9207	7.2262
No log	5.0	280	0.7691	67.6278	56.0075	67.6665	67.5681	7.2787
No log	6.0	336	0.6687	67.8584	56.7482	67.9417	67.8235	7.3213
No log	7.0	392	0.5840	67.9246	56.7	67.9495	67.8923	7.2689
No log	8.0	448	0.5084	69.6732	59.8826	69.6397	69.5534	7.3607
1.4441	9.0	504	0.4686	71.7518	63.1773	71.5904	71.6281	7.0328
1.4441	10.0	560	0.4397	76.241	68.5966	76.1749	76.1173	7.1344
1.4441	11.0	616	0.4121	78.4967	71.125	77.979	78.035	7.341
1.4441	12.0	672	0.3867	81.0066	74.6035	80.9848	80.8444	7.0557
1.4441	13.0	728	0.3682	81.7993	75.5234	81.813	81.6957	7.0918
1.4441	14.0	784	0.3574	82.5172	76.8647	82.5537	82.4686	7.1508
1.4441	15.0	840	0.3449	82.8596	77.533	82.7721	82.6875	7.3377
1.4441	16.0	896	0.3356	83.2423	77.3634	83.0637	83.0135	7.3803
1.4441	17.0	952	0.3274	83.9445	78.4277	83.8606	83.8293	7.3475
0.433	18.0	1008	0.3186	84.8227	79.5215	84.7037	84.6359	7.3705
0.433	19.0	1064	0.3118	85.0225	78.43	84.6229	84.5677	7.3574
0.433	20.0	1120	0.3118	85.0292	79.7869	84.9199	84.867	7.377
0.433	21.0	1176	0.2993	85.5214	80.6124	85.4204	85.3248	7.3738
0.433	22.0	1232	0.3003	86.5962	81.7653	86.5844	86.3981	7.3738
0.433	23.0	1288	0.3015	86.8962	81.7688	86.7715	86.5916	7.3803
0.433	24.0	1344	0.2996	87.1203	82.4587	87.1283	86.9577	7.3869
0.433	25.0	1400	0.2974	87.641	83.2896	87.7111	87.5041	7.3836
0.433	26.0	1456	0.2904	87.7462	83.0436	87.7474	87.5933	7.3541
0.297	27.0	1512	0.2848	88.8773	84.4692	88.8479	88.7003	7.3443
0.297	28.0	1568	0.2932	88.7157	84.409	88.6964	88.6079	7.3639
0.297	29.0	1624	0.2785	88.6821	84.363	88.6386	88.5392	7.3738
0.297	30.0	1680	0.2870	88.7053	84.4901	88.6429	88.5238	7.3738
0.297	31.0	1736	0.2781	89.3913	85.2422	89.367	89.2697	7.3639
0.297	32.0	1792	0.2724	89.3063	85.4257	89.2441	89.1333	7.4066
0.297	33.0	1848	0.2752	89.2602	85.3475	89.2257	89.1599	7.3639
0.297	34.0	1904	0.2703	89.4807	85.8006	89.4182	89.3119	7.3443
0.297	35.0	1960	0.2762	88.9216	84.9247	88.8711	88.8303	7.3803
0.2261	36.0	2016	0.2692	89.4414	85.6311	89.3631	89.2742	7.377
0.2261	37.0	2072	0.2741	89.2527	85.4846	89.2109	89.0969	7.3836
0.2261	38.0	2128	0.2654	89.0749	85.4145	89.0552	88.9593	7.3934
0.2261	39.0	2184	0.2658	88.7595	84.4984	88.7425	88.5844	7.3574
0.2261	40.0	2240	0.2668	89.2718	85.333	89.2728	89.153	7.3541
0.2261	41.0	2296	0.2703	89.1819	85.2561	89.1369	89.0454	7.3672
0.2261	42.0	2352	0.2692	88.8244	84.9209	88.8078	88.6822	7.3311
0.2261	43.0	2408	0.2722	88.8624	85.0424	88.8163	88.6779	7.3738
0.2261	44.0	2464	0.2695	89.7353	85.9085	89.7125	89.6373	7.3836
0.1843	45.0	2520	0.2718	89.5866	85.6088	89.5736	89.4453	7.3541
0.1843	46.0	2576	0.2739	89.289	85.1871	89.2679	89.1394	7.3803
0.1843	47.0	2632	0.2740	89.7117	85.7527	89.6791	89.6127	7.3705
0.1843	48.0	2688	0.2766	89.4973	85.3832	89.485	89.3862	7.3574

Framework versions

• Transformers 4.42.3
• Pytorch 2.3.0+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1