robbert-2023-dutch-large-upos

This model is a fine-tuned version of DTAI-KULeuven/robbert-2023-dutch-large on the universal_dependencies dataset. It achieves the following results on the evaluation set:

• Loss: 0.3606
• Precision: 0.8288
• Recall: 0.7844
• F1: 0.7968
• Accuracy: 0.8898

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: 16
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	1.0	438	0.6318	0.7041	0.6544	0.6603	0.7663
No log	2.0	876	0.5374	0.7741	0.6827	0.7090	0.8075
No log	3.0	1314	0.4318	0.8544	0.7431	0.7527	0.8595
No log	4.0	1752	0.4009	0.8254	0.7677	0.7796	0.8771
No log	5.0	2190	0.3606	0.8288	0.7844	0.7968	0.8898
No log	6.0	2628	0.3700	0.8318	0.8002	0.8108	0.9037
No log	7.0	3066	0.3733	0.8522	0.8024	0.8163	0.9071
No log	8.0	3504	0.3711	0.8659	0.8203	0.8333	0.9189
No log	9.0	3942	0.3846	0.8599	0.8222	0.8343	0.9235
No log	10.0	4380	0.3920	0.8657	0.8263	0.8397	0.9284

Framework versions

• Transformers 4.42.4
• Pytorch 2.3.1+cu121
• Datasets 2.21.0
• Tokenizers 0.19.1