cynthiachan/finetuned-deberta-base-10pct

training

This model is a fine-tuned version of microsoft/deberta-base on the cynthiachan/FeedRef_10pct dataset. It achieves the following results on the evaluation set:

• Loss: 0.0810
• Attackid Precision: 1.0
• Attackid Recall: 1.0
• Attackid F1: 1.0
• Attackid Number: 6
• Cve Precision: 1.0
• Cve Recall: 1.0
• Cve F1: 1.0
• Cve Number: 11
• Defenderthreat Precision: 0.0
• Defenderthreat Recall: 0.0
• Defenderthreat F1: 0.0
• Defenderthreat Number: 2
• Domain Precision: 1.0
• Domain Recall: 0.9565
• Domain F1: 0.9778
• Domain Number: 23
• Email Precision: 1.0
• Email Recall: 1.0
• Email F1: 1.0
• Email Number: 3
• Filepath Precision: 0.8841
• Filepath Recall: 0.8788
• Filepath F1: 0.8815
• Filepath Number: 165
• Hostname Precision: 1.0
• Hostname Recall: 1.0
• Hostname F1: 1.0
• Hostname Number: 12
• Ipv4 Precision: 1.0
• Ipv4 Recall: 1.0
• Ipv4 F1: 1.0
• Ipv4 Number: 12
• Md5 Precision: 0.8333
• Md5 Recall: 0.9615
• Md5 F1: 0.8929
• Md5 Number: 52
• Sha1 Precision: 0.6667
• Sha1 Recall: 0.8571
• Sha1 F1: 0.75
• Sha1 Number: 7
• Sha256 Precision: 0.9565
• Sha256 Recall: 1.0
• Sha256 F1: 0.9778
• Sha256 Number: 44
• Uri Precision: 0.0
• Uri Recall: 0.0
• Uri F1: 0.0
• Uri Number: 1
• Overall Precision: 0.9014
• Overall Recall: 0.9201
• Overall F1: 0.9107
• Overall Accuracy: 0.9851

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: 3.0

Training results

Training Loss	Epoch	Step	Validation Loss	Attackid Precision	Attackid Recall	Attackid F1	Attackid Number	Cve Precision	Cve Recall	Cve F1	Cve Number	Defenderthreat Precision	Defenderthreat Recall	Defenderthreat F1	Defenderthreat Number	Domain Precision	Domain Recall	Domain F1	Domain Number	Email Precision	Email Recall	Email F1	Email Number	Filepath Precision	Filepath Recall	Filepath F1	Filepath Number	Hostname Precision	Hostname Recall	Hostname F1	Hostname Number	Ipv4 Precision	Ipv4 Recall	Ipv4 F1	Ipv4 Number	Md5 Precision	Md5 Recall	Md5 F1	Md5 Number	Sha1 Precision	Sha1 Recall	Sha1 F1	Sha1 Number	Sha256 Precision	Sha256 Recall	Sha256 F1	Sha256 Number	Uri Precision	Uri Recall	Uri F1	Uri Number	Overall Precision	Overall Recall	Overall F1	Overall Accuracy
0.3797	0.37	500	0.1998	0.0	0.0	0.0	6	0.0	0.0	0.0	11	0.0	0.0	0.0	2	0.0286	0.0435	0.0345	23	0.0	0.0	0.0	3	0.5108	0.7152	0.5960	165	0.1774	0.9167	0.2973	12	0.4	0.5	0.4444	12	0.3194	0.4423	0.3710	52	0.0	0.0	0.0	7	0.4588	0.8864	0.6047	44	0.0	0.0	0.0	1	0.3875	0.5858	0.4664	0.9593
0.1713	0.75	1000	0.1619	0.6	0.5	0.5455	6	0.5	0.6364	0.56	11	0.0	0.0	0.0	2	0.6957	0.6957	0.6957	23	0.0	0.0	0.0	3	0.6879	0.6545	0.6708	165	0.5217	1.0	0.6857	12	0.5714	1.0	0.7273	12	0.6667	0.8846	0.7603	52	0.0	0.0	0.0	7	0.7692	0.9091	0.8333	44	0.0	0.0	0.0	1	0.6685	0.7219	0.6942	0.9664
0.1152	1.12	1500	0.1096	0.8333	0.8333	0.8333	6	1.0	1.0	1.0	11	0.0	0.0	0.0	2	0.7826	0.7826	0.7826	23	1.0	1.0	1.0	3	0.7202	0.8424	0.7765	165	1.0	1.0	1.0	12	0.4444	1.0	0.6154	12	0.6944	0.9615	0.8065	52	0.0	0.0	0.0	7	0.8723	0.9318	0.9011	44	0.0	0.0	0.0	1	0.7312	0.8609	0.7908	0.9751
0.1089	1.5	2000	0.1243	1.0	1.0	1.0	6	0.9167	1.0	0.9565	11	0.0	0.0	0.0	2	0.9048	0.8261	0.8636	23	1.0	1.0	1.0	3	0.8011	0.8788	0.8382	165	0.6667	1.0	0.8	12	0.9091	0.8333	0.8696	12	0.7812	0.9615	0.8621	52	0.0	0.0	0.0	7	0.7857	1.0	0.88	44	0.0	0.0	0.0	1	0.8065	0.8876	0.8451	0.9750
0.0947	1.87	2500	0.0913	0.75	1.0	0.8571	6	1.0	1.0	1.0	11	0.0	0.0	0.0	2	0.8462	0.9565	0.8980	23	0.3333	0.6667	0.4444	3	0.8035	0.8424	0.8225	165	0.6	1.0	0.7500	12	1.0	1.0	1.0	12	0.7969	0.9808	0.8793	52	0.0	0.0	0.0	7	0.8302	1.0	0.9072	44	0.0	0.0	0.0	1	0.7952	0.8846	0.8375	0.9792
0.0629	2.25	3000	0.0940	1.0	0.8333	0.9091	6	1.0	1.0	1.0	11	0.0	0.0	0.0	2	0.9565	0.9565	0.9565	23	1.0	1.0	1.0	3	0.8671	0.8303	0.8483	165	1.0	1.0	1.0	12	1.0	1.0	1.0	12	0.9273	0.9808	0.9533	52	0.25	0.1429	0.1818	7	0.8776	0.9773	0.9247	44	0.0	0.0	0.0	1	0.8946	0.8787	0.8866	0.9825
0.0442	2.62	3500	0.1012	1.0	1.0	1.0	6	0.9167	1.0	0.9565	11	0.0	0.0	0.0	2	0.9091	0.8696	0.8889	23	0.75	1.0	0.8571	3	0.8182	0.8727	0.8446	165	1.0	1.0	1.0	12	1.0	1.0	1.0	12	0.92	0.8846	0.9020	52	0.5	1.0	0.6667	7	0.9565	1.0	0.9778	44	0.0	0.0	0.0	1	0.8616	0.9024	0.8815	0.9818
0.0401	3.0	4000	0.0810	1.0	1.0	1.0	6	1.0	1.0	1.0	11	0.0	0.0	0.0	2	1.0	0.9565	0.9778	23	1.0	1.0	1.0	3	0.8841	0.8788	0.8815	165	1.0	1.0	1.0	12	1.0	1.0	1.0	12	0.8333	0.9615	0.8929	52	0.6667	0.8571	0.75	7	0.9565	1.0	0.9778	44	0.0	0.0	0.0	1	0.9014	0.9201	0.9107	0.9851

Framework versions

• Transformers 4.21.2
• Pytorch 1.12.1+cu102
• Datasets 2.4.0
• Tokenizers 0.12.1