scenario-non-kd-po-ner-full-mdeberta_data-univner_half55

This model is a fine-tuned version of haryoaw/scenario-TCR-NER_data-univner_half on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.1503
• Precision: 0.8528
• Recall: 0.8701
• F1: 0.8614
• Accuracy: 0.9848

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: 55
• 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
0.0066	0.5828	500	0.1023	0.8339	0.8461	0.8399	0.9829
0.0066	1.1655	1000	0.0991	0.8499	0.8478	0.8488	0.9838
0.0063	1.7483	1500	0.1052	0.8388	0.8562	0.8474	0.9834
0.005	2.3310	2000	0.1075	0.8267	0.8645	0.8452	0.9830
0.0058	2.9138	2500	0.1067	0.8431	0.8673	0.8550	0.9838
0.0044	3.4965	3000	0.1132	0.8353	0.8605	0.8477	0.9835
0.0042	4.0793	3500	0.1059	0.8429	0.8689	0.8557	0.9842
0.0038	4.6620	4000	0.1047	0.8468	0.8671	0.8569	0.9845
0.0033	5.2448	4500	0.1195	0.8323	0.8557	0.8439	0.9832
0.0037	5.8275	5000	0.1231	0.8376	0.8618	0.8495	0.9831
0.0023	6.4103	5500	0.1207	0.8427	0.8639	0.8532	0.9839
0.0026	6.9930	6000	0.1162	0.8544	0.8566	0.8555	0.9841
0.0029	7.5758	6500	0.1149	0.8418	0.8691	0.8553	0.9841
0.0022	8.1585	7000	0.1342	0.8600	0.8543	0.8571	0.9842
0.0025	8.7413	7500	0.1159	0.8538	0.8566	0.8552	0.9842
0.0023	9.3240	8000	0.1253	0.8468	0.8512	0.8490	0.9837
0.0018	9.9068	8500	0.1288	0.8536	0.8546	0.8541	0.9839
0.0018	10.4895	9000	0.1243	0.8439	0.8523	0.8480	0.9837
0.0015	11.0723	9500	0.1278	0.8546	0.8540	0.8543	0.9839
0.0015	11.6550	10000	0.1337	0.8587	0.8564	0.8576	0.9840
0.0018	12.2378	10500	0.1324	0.8339	0.8562	0.8449	0.9836
0.0017	12.8205	11000	0.1284	0.845	0.8534	0.8492	0.9837
0.0013	13.4033	11500	0.1304	0.8581	0.8602	0.8591	0.9846
0.0015	13.9860	12000	0.1310	0.8335	0.8593	0.8462	0.9836
0.0013	14.5688	12500	0.1356	0.8422	0.8563	0.8492	0.9836
0.0009	15.1515	13000	0.1374	0.8559	0.8713	0.8635	0.9849
0.0013	15.7343	13500	0.1321	0.8399	0.8696	0.8545	0.9844
0.0009	16.3170	14000	0.1412	0.8431	0.8628	0.8528	0.9838
0.0011	16.8998	14500	0.1337	0.8527	0.8645	0.8586	0.9845
0.0006	17.4825	15000	0.1438	0.8620	0.8641	0.8630	0.9847
0.001	18.0653	15500	0.1346	0.8529	0.8683	0.8605	0.9847
0.0005	18.6480	16000	0.1356	0.8620	0.8667	0.8643	0.9850
0.0007	19.2308	16500	0.1426	0.8612	0.8622	0.8617	0.9845
0.0005	19.8135	17000	0.1415	0.8543	0.8588	0.8565	0.9844
0.0007	20.3963	17500	0.1385	0.8590	0.8589	0.8590	0.9844
0.0007	20.9790	18000	0.1405	0.8439	0.8696	0.8565	0.9840
0.0006	21.5618	18500	0.1414	0.8579	0.8648	0.8613	0.9846
0.0005	22.1445	19000	0.1386	0.8579	0.8612	0.8595	0.9845
0.0003	22.7273	19500	0.1447	0.8466	0.8735	0.8598	0.9843
0.0004	23.3100	20000	0.1413	0.8529	0.8626	0.8578	0.9845
0.0003	23.8928	20500	0.1404	0.8580	0.8655	0.8617	0.9850
0.0003	24.4755	21000	0.1466	0.8549	0.8681	0.8615	0.9847
0.0002	25.0583	21500	0.1467	0.8549	0.8645	0.8597	0.9846
0.0001	25.6410	22000	0.1495	0.8497	0.8746	0.8620	0.9847
0.0002	26.2238	22500	0.1479	0.8526	0.8701	0.8613	0.9848
0.0001	26.8065	23000	0.1485	0.8535	0.8655	0.8595	0.9846
0.0002	27.3893	23500	0.1482	0.8535	0.8691	0.8612	0.9847
0.0001	27.9720	24000	0.1501	0.8502	0.8683	0.8592	0.9845
0.0001	28.5548	24500	0.1490	0.8521	0.8670	0.8595	0.9847
0.0001	29.1375	25000	0.1506	0.8503	0.8713	0.8607	0.9847
0.0001	29.7203	25500	0.1503	0.8528	0.8701	0.8614	0.9848

Framework versions

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