README.md

metadata

license: mit
tags:
  - generated_from_trainer
metrics:
  - precision
  - recall
  - f1
  - accuracy
model-index:
  - name: req_mod_ner_modelv2
    results: []
language:
  - nl
widget:
  - text: >-
      De Oplossing ondersteunt het zoeken op de metadata van zaken, documenten
      en objecten en op gegevens uit de basisregistraties die gekoppeld zijn aan
      een zaak.
  - text: >-
      De Oplossing ondersteunt parafering en het plaatsen van een
      gecertificeerde elektronische handtekening.
  - text: >-
      De Aangeboden oplossing stelt de medewerker in staat een zaak te
      registreren.
  - text: >-
      Het Financieel systeem heeft functionaliteit om een
      debiteurenadministratie te voeren.

req_mod_ner_modelv2

This model is a fine-tuned version of pdelobelle/robbert-v2-dutch-ner on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.6964
• Precision: 0.544
• Recall: 0.5862
• F1: 0.5643
• Accuracy: 0.9153

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: 1e-05
• train_batch_size: 2
• eval_batch_size: 2
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 32

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	1.0	120	0.6075	0.8095	0.1466	0.2482	0.8822
No log	2.0	240	0.4917	0.6667	0.1897	0.2953	0.8878
No log	3.0	360	0.4429	0.5	0.3362	0.4021	0.8918
No log	4.0	480	0.4255	0.5	0.4914	0.4957	0.9007
0.507	5.0	600	0.4278	0.5085	0.5172	0.5128	0.9007
0.507	6.0	720	0.4321	0.5294	0.5431	0.5362	0.9064
0.507	7.0	840	0.4574	0.5410	0.5690	0.5546	0.9064
0.507	8.0	960	0.4720	0.5804	0.5603	0.5702	0.9096
0.1626	9.0	1080	0.4947	0.5197	0.5690	0.5432	0.9056
0.1626	10.0	1200	0.5013	0.5159	0.5603	0.5372	0.9096
0.1626	11.0	1320	0.5306	0.5271	0.5862	0.5551	0.9104
0.1626	12.0	1440	0.5450	0.5070	0.6207	0.5581	0.9112
0.0687	13.0	1560	0.5753	0.5152	0.5862	0.5484	0.9112
0.0687	14.0	1680	0.5746	0.5547	0.6121	0.5820	0.9169
0.0687	15.0	1800	0.5925	0.5328	0.6293	0.5771	0.9144
0.0687	16.0	1920	0.6200	0.5656	0.5948	0.5798	0.9144
0.0368	17.0	2040	0.6442	0.5583	0.5776	0.5678	0.9169
0.0368	18.0	2160	0.6468	0.5317	0.5776	0.5537	0.9136
0.0368	19.0	2280	0.6563	0.5403	0.5776	0.5583	0.9153
0.0368	20.0	2400	0.6683	0.5323	0.5690	0.5500	0.9104
0.0227	21.0	2520	0.6766	0.5074	0.5948	0.5476	0.9096
0.0227	22.0	2640	0.6784	0.4965	0.6121	0.5483	0.9072
0.0227	23.0	2760	0.6897	0.5583	0.5776	0.5678	0.9144
0.0227	24.0	2880	0.6858	0.5182	0.6121	0.5613	0.9112
0.0146	25.0	3000	0.6828	0.5224	0.6034	0.5600	0.9128
0.0146	26.0	3120	0.6937	0.5528	0.5862	0.5690	0.9169
0.0146	27.0	3240	0.6939	0.5397	0.5862	0.5620	0.9144
0.0146	28.0	3360	0.6934	0.5476	0.5948	0.5702	0.9169
0.0146	29.0	3480	0.6848	0.5147	0.6034	0.5556	0.9120
0.0132	30.0	3600	0.6864	0.5231	0.5862	0.5528	0.9112
0.0132	31.0	3720	0.6948	0.544	0.5862	0.5643	0.9161
0.0132	32.0	3840	0.6964	0.544	0.5862	0.5643	0.9153

Framework versions

• Transformers 4.24.0
• Pytorch 2.0.0
• Datasets 2.9.0
• Tokenizers 0.11.0