distilbert-base-uncased-finetuned-ner-cadec-no-iob

This model is a fine-tuned version of distilbert-base-uncased on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.4152
• Precision: 0.5826
• Recall: 0.6187
• F1: 0.6001
• Accuracy: 0.9303
• Adr Precision: 0.5109
• Adr Recall: 0.5773
• Adr F1: 0.5421
• Disease Precision: 0.4643
• Disease Recall: 0.4062
• Disease F1: 0.4333
• Drug Precision: 0.8743
• Drug Recall: 0.8889
• Drug F1: 0.8815
• Finding Precision: 0.2143
• Finding Recall: 0.1875
• Finding F1: 0.2000
• Symptom Precision: 0.5556
• Symptom Recall: 0.3448
• Symptom F1: 0.4255
• Macro Avg F1: 0.4965
• Weighted Avg F1: 0.5992

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: 8
• eval_batch_size: 8
• 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	Precision	Recall	F1	Accuracy	Adr Precision	Adr Recall	Adr F1	Disease Precision	Disease Recall	Disease F1	Drug Precision	Drug Recall	Drug F1	Finding Precision	Finding Recall	Finding F1	Symptom Precision	Symptom Recall	Symptom F1	Macro Avg F1	Weighted Avg F1
No log	1.0	125	0.2402	0.4952	0.5462	0.5194	0.9140	0.3893	0.5258	0.4474	0.0	0.0	0.0	0.8883	0.8833	0.8858	0.0	0.0	0.0	0.0	0.0	0.0	0.2666	0.4966
No log	2.0	250	0.2136	0.5380	0.5976	0.5663	0.9239	0.4412	0.5649	0.4955	0.6818	0.4688	0.5556	0.8503	0.8833	0.8665	0.2857	0.0625	0.1026	0.6	0.1034	0.1765	0.4393	0.5573
No log	3.0	375	0.2199	0.5283	0.5660	0.5465	0.9191	0.4492	0.5010	0.4737	0.5455	0.375	0.4444	0.8674	0.8722	0.8698	0.1515	0.1562	0.1538	0.3429	0.4138	0.375	0.4634	0.5492
0.2232	4.0	500	0.2292	0.5622	0.5726	0.5673	0.9228	0.4971	0.5278	0.512	0.0	0.0	0.0	0.8791	0.8889	0.8840	0.1852	0.3125	0.2326	0.4211	0.2759	0.3333	0.3924	0.5601
0.2232	5.0	625	0.2474	0.5863	0.6095	0.5977	0.9265	0.5055	0.5732	0.5372	0.6	0.375	0.4615	0.8785	0.8833	0.8809	0.2273	0.1562	0.1852	0.5333	0.2759	0.3636	0.4857	0.5941
0.2232	6.0	750	0.2474	0.5635	0.5910	0.5769	0.9244	0.4842	0.5381	0.5098	0.375	0.375	0.375	0.8840	0.8889	0.8864	0.16	0.125	0.1404	0.6111	0.3793	0.4681	0.4759	0.5763
0.2232	7.0	875	0.2709	0.5758	0.5963	0.5859	0.9275	0.4991	0.5423	0.5198	0.5	0.2812	0.36	0.8710	0.9	0.8852	0.2683	0.3438	0.3014	0.5385	0.2414	0.3333	0.4799	0.5835
0.0707	8.0	1000	0.2611	0.5752	0.6003	0.5875	0.9282	0.4991	0.5485	0.5226	0.6923	0.2812	0.4	0.8689	0.8833	0.8760	0.2895	0.3438	0.3143	0.4167	0.3448	0.3774	0.4981	0.5870
0.0707	9.0	1125	0.2664	0.5749	0.6227	0.5978	0.9297	0.5111	0.5711	0.5394	0.4375	0.4375	0.4375	0.8710	0.9	0.8852	0.2093	0.2812	0.24	0.5556	0.3448	0.4255	0.5055	0.6003
0.0707	10.0	1250	0.3066	0.5537	0.5778	0.5655	0.9268	0.4761	0.5134	0.4940	0.4545	0.3125	0.3704	0.8610	0.8944	0.8774	0.25	0.2812	0.2647	0.3913	0.3103	0.3462	0.4705	0.5645
0.0707	11.0	1375	0.2980	0.5751	0.5910	0.5830	0.9282	0.4971	0.5340	0.5149	0.4615	0.375	0.4138	0.8602	0.8889	0.8743	0.2917	0.2188	0.25	0.4545	0.3448	0.3922	0.4890	0.5801
0.0293	12.0	1500	0.3272	0.5932	0.6174	0.6050	0.9303	0.5082	0.5732	0.5388	0.6316	0.375	0.4706	0.8901	0.9	0.8950	0.2609	0.1875	0.2182	0.5556	0.3448	0.4255	0.5096	0.6026
0.0293	13.0	1625	0.3161	0.5664	0.6187	0.5914	0.9288	0.4937	0.5691	0.5287	0.3846	0.4688	0.4225	0.8804	0.9	0.8901	0.2308	0.1875	0.2069	0.5	0.3448	0.4082	0.4913	0.5919
0.0293	14.0	1750	0.3529	0.5736	0.6016	0.5873	0.9269	0.4806	0.5361	0.5068	0.5652	0.4062	0.4727	0.8913	0.9111	0.9011	0.3077	0.25	0.2759	0.5238	0.3793	0.44	0.5193	0.5867
0.0293	15.0	1875	0.3381	0.5608	0.6082	0.5835	0.9290	0.5074	0.5649	0.5346	0.2857	0.1875	0.2264	0.8757	0.9	0.8877	0.1731	0.2812	0.2143	0.4167	0.3448	0.3774	0.4481	0.5859
0.0133	16.0	2000	0.3275	0.5833	0.6187	0.6005	0.9307	0.5064	0.5711	0.5368	0.4286	0.375	0.4000	0.8852	0.9	0.8926	0.2759	0.25	0.2623	0.5882	0.3448	0.4348	0.5053	0.6000
0.0133	17.0	2125	0.3623	0.5787	0.6161	0.5968	0.9310	0.4928	0.5649	0.5264	0.6	0.4688	0.5263	0.8852	0.9	0.8926	0.24	0.1875	0.2105	0.5556	0.3448	0.4255	0.5163	0.5962
0.0133	18.0	2250	0.3466	0.5699	0.6187	0.5933	0.9299	0.4937	0.5691	0.5287	0.3889	0.4375	0.4118	0.8901	0.9	0.8950	0.25	0.2188	0.2333	0.5556	0.3448	0.4255	0.4989	0.5944
0.0133	19.0	2375	0.3496	0.5751	0.6214	0.5973	0.9321	0.5101	0.5753	0.5407	0.4	0.375	0.3871	0.8798	0.8944	0.8871	0.1860	0.25	0.2133	0.6875	0.3793	0.4889	0.5034	0.6007
0.0075	20.0	2500	0.3676	0.5898	0.6280	0.6083	0.9314	0.5090	0.5814	0.5428	0.5185	0.4375	0.4746	0.8852	0.9	0.8926	0.2692	0.2188	0.2414	0.6471	0.3793	0.4783	0.5259	0.6078
0.0075	21.0	2625	0.3658	0.5816	0.6253	0.6027	0.9306	0.4991	0.5753	0.5345	0.5185	0.4375	0.4746	0.8811	0.9056	0.8932	0.2593	0.2188	0.2373	0.6471	0.3793	0.4783	0.5236	0.6024
0.0075	22.0	2750	0.3803	0.5804	0.6187	0.5990	0.9294	0.5148	0.5753	0.5433	0.3846	0.3125	0.3448	0.8859	0.9056	0.8956	0.2059	0.2188	0.2121	0.4545	0.3448	0.3922	0.4776	0.5988
0.0075	23.0	2875	0.3795	0.5954	0.6174	0.6062	0.9305	0.5139	0.5711	0.5410	0.5652	0.4062	0.4727	0.8852	0.9	0.8926	0.2609	0.1875	0.2182	0.5556	0.3448	0.4255	0.5100	0.6036
0.0051	24.0	3000	0.3849	0.5774	0.6148	0.5955	0.9295	0.5093	0.5670	0.5366	0.4444	0.375	0.4068	0.8798	0.8944	0.8871	0.2121	0.2188	0.2154	0.4583	0.3793	0.4151	0.4922	0.5961
0.0051	25.0	3125	0.3847	0.5911	0.6293	0.6096	0.9303	0.5247	0.5918	0.5562	0.4828	0.4375	0.4590	0.875	0.8944	0.8846	0.1724	0.1562	0.1639	0.5556	0.3448	0.4255	0.4979	0.6085
0.0051	26.0	3250	0.3917	0.5901	0.6266	0.6078	0.9298	0.5165	0.5794	0.5462	0.4667	0.4375	0.4516	0.8804	0.9	0.8901	0.2759	0.25	0.2623	0.5556	0.3448	0.4255	0.5151	0.6072
0.0051	27.0	3375	0.3915	0.5901	0.6306	0.6097	0.9306	0.5182	0.5876	0.5507	0.4828	0.4375	0.4590	0.8852	0.9	0.8926	0.2414	0.2188	0.2295	0.5263	0.3448	0.4167	0.5097	0.6093
0.0034	28.0	3500	0.4010	0.5881	0.6253	0.6061	0.9305	0.5240	0.5856	0.5531	0.4167	0.3125	0.3571	0.8757	0.9	0.8877	0.2162	0.25	0.2319	0.5556	0.3448	0.4255	0.4911	0.6058
0.0034	29.0	3625	0.4136	0.5955	0.6293	0.6119	0.9313	0.5212	0.5835	0.5506	0.4828	0.4375	0.4590	0.8859	0.9056	0.8956	0.2692	0.2188	0.2414	0.5263	0.3448	0.4167	0.5127	0.6105
0.0034	30.0	3750	0.4072	0.5918	0.6293	0.6100	0.9312	0.5191	0.5876	0.5513	0.4615	0.375	0.4138	0.8804	0.9	0.8901	0.2581	0.25	0.2540	0.625	0.3448	0.4444	0.5107	0.6093
0.0034	31.0	3875	0.4081	0.5995	0.6240	0.6115	0.9307	0.5294	0.5753	0.5514	0.4375	0.4375	0.4375	0.8804	0.9	0.8901	0.32	0.25	0.2807	0.4762	0.3448	0.4000	0.5119	0.6098
0.0025	32.0	4000	0.4022	0.5885	0.6319	0.6094	0.9312	0.5152	0.5938	0.5517	0.5185	0.4375	0.4746	0.875	0.8944	0.8846	0.25	0.1875	0.2143	0.5	0.3448	0.4082	0.5067	0.6078
0.0025	33.0	4125	0.4066	0.5821	0.6266	0.6036	0.9312	0.5108	0.5876	0.5465	0.4643	0.4062	0.4333	0.8743	0.8889	0.8815	0.2414	0.2188	0.2295	0.5556	0.3448	0.4255	0.5033	0.6033
0.0025	34.0	4250	0.4049	0.5865	0.6306	0.6078	0.9318	0.5198	0.5959	0.5552	0.4815	0.4062	0.4407	0.8696	0.8889	0.8791	0.2	0.1875	0.1935	0.5556	0.3448	0.4255	0.4988	0.6071
0.0025	35.0	4375	0.4129	0.5741	0.6187	0.5956	0.9294	0.5009	0.5773	0.5364	0.5	0.4375	0.4667	0.8689	0.8833	0.8760	0.2069	0.1875	0.1967	0.5556	0.3448	0.4255	0.5003	0.5955
0.002	36.0	4500	0.4134	0.5843	0.6266	0.6047	0.9303	0.5117	0.5876	0.5470	0.5	0.4375	0.4667	0.8743	0.8889	0.8815	0.2222	0.1875	0.2034	0.5556	0.3448	0.4255	0.5048	0.6039
0.002	37.0	4625	0.4138	0.5828	0.6266	0.6039	0.9303	0.5099	0.5856	0.5451	0.4815	0.4062	0.4407	0.875	0.8944	0.8846	0.2414	0.2188	0.2295	0.5556	0.3448	0.4255	0.5051	0.6034
0.002	38.0	4750	0.4126	0.5804	0.6187	0.5990	0.9297	0.5100	0.5794	0.5425	0.4444	0.375	0.4068	0.8743	0.8889	0.8815	0.2069	0.1875	0.1967	0.5556	0.3448	0.4255	0.4906	0.5982
0.002	39.0	4875	0.4139	0.5797	0.6187	0.5986	0.9301	0.5118	0.5794	0.5435	0.4286	0.375	0.4000	0.8743	0.8889	0.8815	0.1935	0.1875	0.1905	0.5556	0.3448	0.4255	0.4882	0.5983
0.0017	40.0	5000	0.4152	0.5826	0.6187	0.6001	0.9303	0.5109	0.5773	0.5421	0.4643	0.4062	0.4333	0.8743	0.8889	0.8815	0.2143	0.1875	0.2000	0.5556	0.3448	0.4255	0.4965	0.5992

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.0+cu121
• Datasets 2.15.0
• Tokenizers 0.15.0