NorBERT3-Large_FGN

This model is a fine-tuned version of ltg/norbert3-large on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.4464
• F1-score: 0.8361

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: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	F1-score
No log	1.0	120	0.4704	0.8059
No log	2.0	240	0.4812	0.8151
No log	3.0	360	0.8574	0.7980
No log	4.0	480	1.0347	0.7849
0.4487	5.0	600	1.3767	0.8214
0.4487	6.0	720	1.4754	0.8159
0.4487	7.0	840	1.4435	0.8333
0.4487	8.0	960	1.3504	0.8338
0.0369	9.0	1080	1.3592	0.8338
0.0369	10.0	1200	1.3162	0.8299
0.0369	11.0	1320	1.3533	0.8317
0.0369	12.0	1440	1.3645	0.8317
0.0093	13.0	1560	1.3810	0.8299
0.0093	14.0	1680	1.4039	0.8317
0.0093	15.0	1800	1.4163	0.8289
0.0093	16.0	1920	1.4252	0.8299
0.0057	17.0	2040	1.4199	0.8299
0.0057	18.0	2160	1.4326	0.8317
0.0057	19.0	2280	1.4467	0.8256
0.0057	20.0	2400	1.4464	0.8361

Framework versions

• Transformers 4.41.2
• Pytorch 2.3.0+cu121
• Datasets 2.19.2
• Tokenizers 0.19.1