distilbert-base-uncased_fold_9_ternary_v1

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

• Loss: 1.9406
• F1: 0.7841

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: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 25

Training results

Training Loss	Epoch	Step	Validation Loss	F1
No log	1.0	292	0.5684	0.7635
0.5656	2.0	584	0.5753	0.7725
0.5656	3.0	876	0.6159	0.7866
0.2499	4.0	1168	0.7743	0.7828
0.2499	5.0	1460	0.9820	0.7674
0.1153	6.0	1752	1.2383	0.7738
0.0547	7.0	2044	1.2468	0.7815
0.0547	8.0	2336	1.3480	0.7622
0.0233	9.0	2628	1.3791	0.7892
0.0233	10.0	2920	1.4344	0.7841
0.0142	11.0	3212	1.4958	0.7802
0.0087	12.0	3504	1.5714	0.7674
0.0087	13.0	3796	1.6129	0.7956
0.0111	14.0	4088	1.7799	0.7751
0.0111	15.0	4380	1.7272	0.7789
0.0055	16.0	4672	1.7696	0.7866
0.0055	17.0	4964	1.8622	0.7789
0.003	18.0	5256	1.8563	0.7802
0.0004	19.0	5548	1.8993	0.7815
0.0004	20.0	5840	1.9199	0.7853
0.0005	21.0	6132	1.9003	0.7879
0.0005	22.0	6424	1.9161	0.7828
0.0011	23.0	6716	1.9691	0.7815
0.0017	24.0	7008	1.9492	0.7841
0.0017	25.0	7300	1.9406	0.7841

Framework versions

• Transformers 4.21.0
• Pytorch 1.12.0+cu113
• Datasets 2.4.0
• Tokenizers 0.12.1