MeMo_BERT-SA_ScandiBERT

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

• Loss: 2.1392
• F1-score: 0.7266

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	297	0.8686	0.6373
0.8175	2.0	594	0.8961	0.7140
0.8175	3.0	891	0.8175	0.7237
0.5748	4.0	1188	1.2192	0.7226
0.5748	5.0	1485	1.1315	0.7003
0.3963	6.0	1782	1.3036	0.6791
0.2579	7.0	2079	1.2430	0.7100
0.2579	8.0	2376	1.7559	0.7143
0.1768	9.0	2673	1.8862	0.7070
0.1768	10.0	2970	1.9733	0.7137
0.1065	11.0	3267	2.2073	0.7071
0.0559	12.0	3564	2.1912	0.7184
0.0559	13.0	3861	2.1392	0.7266
0.0367	14.0	4158	2.3618	0.6950
0.0367	15.0	4455	2.3324	0.7223
0.026	16.0	4752	2.4711	0.7108
0.0041	17.0	5049	2.5320	0.7108
0.0041	18.0	5346	2.6578	0.6948
0.0059	19.0	5643	2.6631	0.6991
0.0059	20.0	5940	2.6313	0.7067

Framework versions

• Transformers 4.38.2
• Pytorch 2.2.1+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2