ner_column_TQ

This model is a fine-tuned version of Gladiator/microsoft-deberta-v3-large_ner_conll2003 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.1949
• Precision: 0.8546
• Recall: 0.8533
• F1: 0.8540
• Accuracy: 0.9154

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: 64
• eval_batch_size: 64
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	1.0	702	0.2342	0.7774	0.7496	0.7632	0.8833
0.369	2.0	1404	0.1708	0.8050	0.8048	0.8049	0.9033
0.1681	3.0	2106	0.1646	0.8007	0.8078	0.8043	0.9054
0.1681	4.0	2808	0.1469	0.8250	0.8335	0.8292	0.9133
0.14	5.0	3510	0.1465	0.8235	0.8345	0.8290	0.9137
0.1279	6.0	4212	0.1517	0.8165	0.8323	0.8244	0.9127
0.1279	7.0	4914	0.1474	0.8224	0.8370	0.8297	0.9138
0.1212	8.0	5616	0.1500	0.8255	0.8409	0.8331	0.9141
0.1165	9.0	6318	0.1545	0.8297	0.8390	0.8343	0.9142
0.1138	10.0	7020	0.1590	0.8342	0.8467	0.8404	0.9150
0.1138	11.0	7722	0.1588	0.8383	0.8474	0.8428	0.9156
0.1099	12.0	8424	0.1547	0.8425	0.8446	0.8435	0.9156
0.1071	13.0	9126	0.1565	0.8475	0.8471	0.8473	0.9164
0.1071	14.0	9828	0.1625	0.8440	0.8489	0.8464	0.9156
0.1031	15.0	10530	0.1680	0.8486	0.8510	0.8498	0.9160
0.0992	16.0	11232	0.1722	0.8529	0.8505	0.8517	0.9156
0.0992	17.0	11934	0.1771	0.8527	0.8529	0.8528	0.9159
0.094	18.0	12636	0.1862	0.8555	0.8531	0.8543	0.9159
0.0892	19.0	13338	0.1884	0.8534	0.8534	0.8534	0.9156
0.086	20.0	14040	0.1949	0.8546	0.8533	0.8540	0.9154

Framework versions

• Transformers 4.33.2
• Pytorch 2.0.1+cu117
• Datasets 2.14.5
• Tokenizers 0.13.3