distilbert-base-uncased_fold_3_binary_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: 1.9405
• F1: 0.7878

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	289	0.4630	0.7897
0.3954	2.0	578	0.4549	0.7936
0.3954	3.0	867	0.6527	0.7868
0.1991	4.0	1156	0.7510	0.7951
0.1991	5.0	1445	0.9327	0.8000
0.095	6.0	1734	1.0974	0.7859
0.0347	7.0	2023	1.2692	0.7919
0.0347	8.0	2312	1.3718	0.7921
0.0105	9.0	2601	1.4679	0.7999
0.0105	10.0	2890	1.5033	0.8070
0.0079	11.0	3179	1.6074	0.8008
0.0079	12.0	3468	1.6921	0.7904
0.0053	13.0	3757	1.7079	0.7945
0.0054	14.0	4046	1.8361	0.7887
0.0054	15.0	4335	1.7695	0.7873
0.0046	16.0	4624	1.7934	0.7917
0.0046	17.0	4913	1.8036	0.8008
0.0064	18.0	5202	1.8780	0.7888
0.0064	19.0	5491	1.8943	0.7923
0.0032	20.0	5780	1.8694	0.7905
0.002	21.0	6069	1.9348	0.7869
0.002	22.0	6358	1.9578	0.7804
0.0036	23.0	6647	1.9438	0.7827
0.0036	24.0	6936	1.9386	0.7878
0.0011	25.0	7225	1.9405	0.7878

Framework versions

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