scenario-non-kd-scr-ner-half-mdeberta_data-univner_full44

This model is a fine-tuned version of microsoft/mdeberta-v3-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.3028
• Precision: 0.6277
• Recall: 0.5869
• F1: 0.6066
• Accuracy: 0.9615

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: 3e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 44
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
0.36	0.2910	500	0.2872	0.2875	0.1173	0.1666	0.9284
0.2389	0.5821	1000	0.2086	0.3476	0.2561	0.2949	0.9388
0.1727	0.8731	1500	0.1810	0.4363	0.3748	0.4033	0.9465
0.1338	1.1641	2000	0.1644	0.4626	0.4675	0.4650	0.9514
0.1122	1.4552	2500	0.1560	0.4983	0.5063	0.5023	0.9538
0.1033	1.7462	3000	0.1504	0.5354	0.5128	0.5238	0.9564
0.0919	2.0373	3500	0.1475	0.5073	0.5452	0.5256	0.9558
0.0665	2.3283	4000	0.1547	0.5536	0.5578	0.5557	0.9583
0.066	2.6193	4500	0.1513	0.5345	0.5760	0.5545	0.9581
0.0642	2.9104	5000	0.1489	0.5750	0.5683	0.5716	0.9605
0.0495	3.2014	5500	0.1597	0.5769	0.5711	0.5740	0.9600
0.0437	3.4924	6000	0.1643	0.5848	0.5680	0.5763	0.9603
0.0444	3.7835	6500	0.1615	0.5884	0.5898	0.5891	0.9607
0.0409	4.0745	7000	0.1723	0.5869	0.5761	0.5815	0.9606
0.0313	4.3655	7500	0.1740	0.5871	0.5930	0.5900	0.9606
0.032	4.6566	8000	0.1682	0.5911	0.6031	0.5971	0.9611
0.0304	4.9476	8500	0.1771	0.6070	0.5783	0.5923	0.9613
0.0228	5.2386	9000	0.1843	0.5817	0.6045	0.5929	0.9608
0.0216	5.5297	9500	0.1841	0.5938	0.6142	0.6038	0.9609
0.0232	5.8207	10000	0.1957	0.5816	0.5998	0.5906	0.9600
0.0201	6.1118	10500	0.1982	0.6049	0.5963	0.6006	0.9611
0.0153	6.4028	11000	0.2040	0.5919	0.6057	0.5987	0.9602
0.0165	6.6938	11500	0.2039	0.6000	0.5988	0.5994	0.9609
0.0165	6.9849	12000	0.2076	0.5963	0.5913	0.5938	0.9606
0.0121	7.2759	12500	0.2178	0.6015	0.5833	0.5923	0.9604
0.012	7.5669	13000	0.2186	0.6206	0.5902	0.6050	0.9613
0.0126	7.8580	13500	0.2218	0.5882	0.6191	0.6033	0.9600
0.0098	8.1490	14000	0.2296	0.6164	0.5911	0.6035	0.9617
0.0091	8.4400	14500	0.2332	0.5986	0.5976	0.5981	0.9607
0.0097	8.7311	15000	0.2322	0.6053	0.5996	0.6024	0.9613
0.0089	9.0221	15500	0.2355	0.6174	0.6034	0.6103	0.9612
0.0064	9.3132	16000	0.2440	0.6306	0.5835	0.6061	0.9614
0.0071	9.6042	16500	0.2451	0.6220	0.5761	0.5982	0.9609
0.0073	9.8952	17000	0.2461	0.6203	0.5990	0.6095	0.9616
0.0064	10.1863	17500	0.2506	0.6213	0.5900	0.6052	0.9615
0.005	10.4773	18000	0.2547	0.6226	0.5970	0.6096	0.9617
0.0058	10.7683	18500	0.2553	0.6374	0.5897	0.6126	0.9620
0.0054	11.0594	19000	0.2624	0.6232	0.5840	0.6030	0.9617
0.0044	11.3504	19500	0.2655	0.6262	0.5946	0.6100	0.9620
0.0048	11.6414	20000	0.2654	0.6154	0.5989	0.6070	0.9616
0.0042	11.9325	20500	0.2724	0.6306	0.5806	0.6046	0.9616
0.004	12.2235	21000	0.2707	0.6052	0.5920	0.5985	0.9607
0.0035	12.5146	21500	0.2714	0.5962	0.5986	0.5974	0.9607
0.0041	12.8056	22000	0.2755	0.6263	0.5858	0.6053	0.9616
0.0035	13.0966	22500	0.2842	0.6350	0.5814	0.6071	0.9614
0.0033	13.3877	23000	0.2763	0.6317	0.5868	0.6084	0.9614
0.0028	13.6787	23500	0.2831	0.6141	0.5976	0.6057	0.9616
0.0031	13.9697	24000	0.2797	0.6141	0.6064	0.6102	0.9614
0.0024	14.2608	24500	0.2873	0.5980	0.6038	0.6009	0.9611
0.0025	14.5518	25000	0.2913	0.6055	0.5980	0.6017	0.9612
0.003	14.8428	25500	0.2885	0.6208	0.5843	0.6020	0.9615
0.0023	15.1339	26000	0.2923	0.6255	0.5849	0.6045	0.9618
0.0019	15.4249	26500	0.2875	0.6221	0.6015	0.6116	0.9619
0.0027	15.7159	27000	0.2898	0.6241	0.5967	0.6101	0.9619
0.0024	16.0070	27500	0.2943	0.6146	0.5895	0.6018	0.9612
0.0016	16.2980	28000	0.2996	0.6199	0.5928	0.6060	0.9614
0.0022	16.5891	28500	0.3028	0.6277	0.5869	0.6066	0.9615

Framework versions

• Transformers 4.44.2
• Pytorch 2.1.1+cu121
• Datasets 2.14.5
• Tokenizers 0.19.1