manifesto-dutch-binary-relevance

This model is a fine-tuned version of pdelobelle/robbert-v2-dutch-base.

Example usage

from transformers import pipeline

pipe = pipeline("text-classification", 
                model="joris/manifesto-dutch-binary-relevance",
                trust_remote_code=True)

print(pipe("De digitale versie lees je op d66.nl/verkiezingsprogramma")) 
print(pipe("Duizenden studenten, net afgestudeerden en starters hebben op dit moment geen zicht op een (betaalbare) woning."))


## [{'label': 'LABEL_1', 'score': 0.9609444737434387}] # is 000
## [{'label': 'LABEL_0', 'score': 0.9993253946304321}] # some other code

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

	Precision	Recall	F1-Score	Support
0	0.98	0.99	0.99	10043
1	0.88	0.76	0.82	714
Accuracy			0.98	10757
Macro avg	0.93	0.88	0.90	10757
Weighted avg	0.98	0.98	0.98	10757

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• optimizer: {'name': 'AdamW', 'weight_decay': 0.004, 'clipnorm': None, 'global_clipnorm': None, 'clipvalue': None, 'use_ema': False, 'ema_momentum': 0.99, 'ema_overwrite_frequency': None, 'jit_compile': True, 'is_legacy_optimizer': False, 'learning_rate': 2e-05, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-07, 'amsgrad': False}
• training_precision: float32

Training results

Framework versions

• Transformers 4.34.1
• TensorFlow 2.14.0
• Tokenizers 0.14.1