gbert-large-upos

This model is a fine-tuned version of deepset/gbert-large on the universal_dependencies dataset. It achieves the following results on the evaluation set:

• Loss: 0.1996
• Precision: 0.8253
• Recall: 0.7827
• F1: 0.7912
• Accuracy: 0.9414

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	0.3197	0.8098	0.7291	0.7486	0.8936
No log	2.0	876	0.2261	0.8287	0.7679	0.7832	0.9269
No log	3.0	1314	0.1996	0.8253	0.7827	0.7912	0.9414
No log	4.0	1752	0.2183	0.8162	0.8006	0.8041	0.9435
No log	5.0	2190	0.2120	0.8198	0.8025	0.8074	0.9496
No log	6.0	2628	0.2339	0.8207	0.8068	0.8116	0.9489
No log	7.0	3066	0.2728	0.8156	0.8045	0.8071	0.9486
No log	8.0	3504	0.2790	0.8205	0.8110	0.8132	0.9527
No log	9.0	3942	0.2854	0.8306	0.8096	0.8146	0.9527
No log	10.0	4380	0.2906	0.8299	0.8115	0.8151	0.9534

Framework versions

• Transformers 4.42.4
• Pytorch 2.4.0+cu121
• Datasets 2.21.0
• Tokenizers 0.19.1