NL_BERT_michelin_finetuned

This model is a fine-tuned version of GroNLP/bert-base-dutch-cased on a Dutch restaurant reviews dataset. Provide Dutch review text to the API on the right and receive a score that indicates whether this restaurant is eligible for a Michelin star ;) It achieves the following results on the evaluation set:

• Loss: 0.0637
• Accuracy: 0.9836
• Recall: 0.5486
• Precision: 0.7914
• F1: 0.6480
• Mse: 0.0164

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: 5e-05
• train_batch_size: 32
• eval_batch_size: 128
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Recall	Precision	F1	Mse
0.1043	1.0	3647	0.0961	0.9792	0.3566	0.7606	0.4856	0.0208
0.0799	2.0	7294	0.0797	0.9803	0.4364	0.7415	0.5495	0.0197
0.0589	3.0	10941	0.0637	0.9836	0.5486	0.7914	0.6480	0.0164

Framework versions

• Transformers 4.18.0
• Pytorch 1.11.0+cu113
• Datasets 2.1.0
• Tokenizers 0.12.1