nb-bert-edu-scorer

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

• Loss: 0.7249
• Precision: 0.3908
• Recall: 0.3347
• F1 Macro: 0.3334
• Accuracy: 0.48

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: 3e-05
• train_batch_size: 256
• eval_batch_size: 128
• seed: 0
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1 Macro	Accuracy
No log	0	0	2.4712	0.0986	0.1654	0.0877	0.3496
0.7734	2.6882	1000	0.7629	0.3988	0.3258	0.3215	0.4652
0.7602	5.3763	2000	0.7505	0.3938	0.3319	0.3284	0.4584
0.7548	8.0645	3000	0.7345	0.3924	0.3345	0.3319	0.4758
0.731	10.7527	4000	0.7300	0.3951	0.3359	0.3335	0.4756
0.7481	13.4409	5000	0.7274	0.3957	0.3356	0.3337	0.4818
0.7255	16.1290	6000	0.7263	0.3910	0.3361	0.3339	0.4754
0.7371	18.8172	7000	0.7249	0.3908	0.3347	0.3334	0.48

Framework versions

• Transformers 4.53.2
• Pytorch 2.7.1+cu126
• Datasets 4.0.0
• Tokenizers 0.21.2