README.md

metadata

license: mit
base_model: microsoft/deberta-v3-base
tags:
  - generated_from_trainer
datasets:
  - maccrobat_biomedical_ner
metrics:
  - precision
  - recall
  - f1
  - accuracy
model-index:
  - name: deberta-v3-base-finetuned-ner
    results:
      - task:
          name: Token Classification
          type: token-classification
        dataset:
          name: maccrobat_biomedical_ner
          type: maccrobat_biomedical_ner
          config: default
          split: train
          args: default
        metrics:
          - name: Precision
            type: precision
            value: 0.7843711467324291
          - name: Recall
            type: recall
            value: 0.7816003686069728
          - name: F1
            type: f1
            value: 0.7829833064081853
          - name: Accuracy
            type: accuracy
            value: 0.8584199081903842

deberta-v3-base-finetuned-ner

This model is a fine-tuned version of microsoft/deberta-v3-base on the maccrobat_biomedical_ner dataset. It achieves the following results on the evaluation set:

• Loss: 0.9704
• Precision: 0.7844
• Recall: 0.7816
• F1: 0.7830
• Accuracy: 0.8584

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: 4.555607052152088e-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: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	1.0	20	0.9499	0.7670	0.7685	0.7678	0.8477
No log	2.0	40	0.9042	0.7721	0.7629	0.7675	0.8484
No log	3.0	60	0.9360	0.7674	0.7573	0.7623	0.8475
No log	4.0	80	0.8984	0.7630	0.7589	0.7609	0.8442
No log	5.0	100	0.8159	0.7695	0.7701	0.7698	0.8495
No log	6.0	120	0.8086	0.7557	0.7730	0.7643	0.8454
No log	7.0	140	0.7937	0.7766	0.7712	0.7739	0.8509
No log	8.0	160	0.8430	0.7703	0.7707	0.7705	0.8513
No log	9.0	180	0.8711	0.7715	0.7710	0.7712	0.8517
No log	10.0	200	0.8649	0.7687	0.7626	0.7656	0.8485
No log	11.0	220	0.8686	0.7817	0.7635	0.7725	0.8516
No log	12.0	240	0.8644	0.7765	0.7802	0.7784	0.8546
No log	13.0	260	0.8680	0.7771	0.7796	0.7783	0.8550
No log	14.0	280	0.8845	0.7728	0.7748	0.7738	0.8528
No log	15.0	300	0.9084	0.7774	0.7713	0.7743	0.8537
No log	16.0	320	0.9396	0.7782	0.7659	0.7720	0.8509
No log	17.0	340	0.9338	0.7776	0.7781	0.7778	0.8547
No log	18.0	360	0.9205	0.7749	0.7770	0.7759	0.8537
No log	19.0	380	0.9426	0.7781	0.7724	0.7752	0.8523
No log	20.0	400	0.9403	0.7769	0.7827	0.7798	0.8550
No log	21.0	420	0.9393	0.7795	0.7713	0.7754	0.8536
No log	22.0	440	0.9618	0.7771	0.7790	0.7780	0.8547
No log	23.0	460	0.9420	0.7814	0.7836	0.7825	0.8582
No log	24.0	480	0.9455	0.7842	0.7808	0.7825	0.8583
0.0412	25.0	500	0.9599	0.7821	0.7801	0.7811	0.8571
0.0412	26.0	520	0.9518	0.7815	0.7833	0.7824	0.8578
0.0412	27.0	540	0.9570	0.7800	0.7818	0.7809	0.8567
0.0412	28.0	560	0.9634	0.7819	0.7801	0.7810	0.8573
0.0412	29.0	580	0.9685	0.7818	0.7831	0.7825	0.8579
0.0412	30.0	600	0.9704	0.7844	0.7816	0.7830	0.8584

Framework versions

• Transformers 4.39.3
• Pytorch 2.2.1
• Datasets 2.18.0
• Tokenizers 0.15.2