VaccinChatSentenceClassifierDutch_fromBERTje2

This model is a fine-tuned version of GroNLP/bert-base-dutch-cased on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.5112
• Accuracy: 0.9004

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: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-06
• lr_scheduler_type: linear
• num_epochs: 15.0

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
4.1505	1.0	1320	3.3293	0.3793
2.7333	2.0	2640	2.2295	0.6133
2.0189	3.0	3960	1.5134	0.7587
1.2504	4.0	5280	1.0765	0.8282
0.7733	5.0	6600	0.7937	0.8629
0.5217	6.0	7920	0.6438	0.8784
0.3148	7.0	9240	0.5733	0.8857
0.2067	8.0	10560	0.5362	0.8912
0.1507	9.0	11880	0.5098	0.8903
0.1024	10.0	13200	0.5078	0.8976
0.0837	11.0	14520	0.5054	0.8967
0.0608	12.0	15840	0.5062	0.8958
0.0426	13.0	17160	0.5072	0.9013
0.0374	14.0	18480	0.5110	0.9040
0.0346	15.0	19800	0.5112	0.9004

Framework versions

• Transformers 4.13.0.dev0
• Pytorch 1.10.0
• Datasets 1.16.1
• Tokenizers 0.10.3