scenario-non-kd-po-ner-full-xlmr_data-univner_half66

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: 0.1331
• Precision: 0.8406
• Recall: 0.8491
• F1: 0.8448
• Accuracy: 0.9836

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: 66
• 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
0.0102	0.5828	500	0.0979	0.8297	0.8550	0.8422	0.9832
0.0109	1.1655	1000	0.0889	0.8346	0.8466	0.8406	0.9835
0.0084	1.7483	1500	0.0932	0.8491	0.8462	0.8477	0.9839
0.0075	2.3310	2000	0.0919	0.8434	0.8437	0.8436	0.9835
0.0072	2.9138	2500	0.1043	0.8278	0.8380	0.8329	0.9826
0.0058	3.4965	3000	0.1020	0.8370	0.8468	0.8419	0.9832
0.0059	4.0793	3500	0.1030	0.8467	0.8458	0.8463	0.9839
0.005	4.6620	4000	0.1182	0.8492	0.8326	0.8408	0.9830
0.005	5.2448	4500	0.1141	0.8235	0.8510	0.8370	0.9821
0.0044	5.8275	5000	0.1184	0.8273	0.8572	0.8420	0.9828
0.0045	6.4103	5500	0.1213	0.8417	0.8494	0.8455	0.9833
0.0048	6.9930	6000	0.1126	0.8413	0.8413	0.8413	0.9835
0.0043	7.5758	6500	0.1240	0.8363	0.8472	0.8417	0.9830
0.0036	8.1585	7000	0.1185	0.8470	0.8450	0.8460	0.9838
0.0032	8.7413	7500	0.1249	0.8338	0.8391	0.8365	0.9828
0.0022	9.3240	8000	0.1260	0.8351	0.8499	0.8425	0.9835
0.003	9.9068	8500	0.1208	0.8273	0.8420	0.8346	0.9827
0.0024	10.4895	9000	0.1216	0.8451	0.8463	0.8457	0.9836
0.0027	11.0723	9500	0.1214	0.8410	0.8390	0.8400	0.9832
0.0019	11.6550	10000	0.1234	0.8419	0.8458	0.8438	0.9833
0.0023	12.2378	10500	0.1289	0.8339	0.8502	0.8420	0.9830
0.0019	12.8205	11000	0.1286	0.8400	0.8401	0.8401	0.9832
0.002	13.4033	11500	0.1331	0.8406	0.8491	0.8448	0.9836

Framework versions

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