ANGKJ1995/generic-course-prediction

my_awesome_model

This model is a fine-tuned version of distilbert/distilbert-base-uncased on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.4282
• F1: 0.6908

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: 2e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	F1
0.1734	1.0	1685	0.1698	0.5775
0.1401	2.0	3370	0.1750	0.6176
0.1171	3.0	5055	0.1967	0.6104
0.082	4.0	6740	0.2097	0.6597
0.0571	5.0	8425	0.2515	0.6580
0.0611	6.0	10110	0.2643	0.6648
0.0415	7.0	11795	0.3226	0.6473
0.0341	8.0	13480	0.2959	0.6786
0.0311	9.0	15165	0.3864	0.6736
0.0223	10.0	16850	0.3742	0.6739
0.0229	11.0	18535	0.3545	0.6646
0.0204	12.0	20220	0.3989	0.6738
0.0198	13.0	21905	0.3672	0.6798
0.0128	14.0	23590	0.4160	0.6750
0.0128	15.0	25275	0.4282	0.6908
0.0092	16.0	26960	0.4178	0.6814
0.0143	17.0	28645	0.4289	0.6754
0.0076	18.0	30330	0.4366	0.682
0.0071	19.0	32015	0.4441	0.6867
0.0074	20.0	33700	0.4530	0.6874

Framework versions

• Transformers 4.57.3
• Pytorch 2.9.0+cu126
• Datasets 4.0.0
• Tokenizers 0.22.1