bert-unformatted-network-data-test-6-types

This model is a fine-tuned version of roberta-large on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.1318
• F1: 0.9624

EXAMPLE FULL NAMES:

label_0 = UDP-lag DDoS, label_1 = benign, label_2 = SYN flood, label_3 = NetBIOS, label_4 = MSSQL, label_5 = LDAP

1. Benign traffic from training data
2. Benign traffic from outside training data
3. malicious UDP-Lag DDoS attack from training data
4. malicious UDP-Lag DDoS attack from outside of training data
5. malicious SYN flood attack from training data
6. malicious SYN flood attack from outside of training data
7. malicious NetBIOS DDoS attack from training data
8. malicious NetBIOS DDoS attack from outside of training data
9. malicious MSSQL DDoS attack from training data
10. malicious MSSQL DDoS attack from outside of training data
11. malicious LDAP DDoS attack from training data
12. malicious LDAP DDoS attack from outside of training data

examples from CIC-DDoS2019 (formatted for model training) https://colab.research.google.com/drive/1PmLep9D3NfMhYsX0soTBhfVXFkawGgGx?authuser=0#scrollTo=ReaH6NCljdsn

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: 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: 5

Training results

Training Loss	Epoch	Step	Validation Loss	F1
0.1427	1.0	2250	0.1279	0.9622
0.1348	2.0	4500	0.1517	0.9624
0.1331	3.0	6750	0.1467	0.9613
0.1407	4.0	9000	0.1294	0.9623
0.1229	5.0	11250	0.1318	0.9624

Framework versions

• Transformers 4.42.0.dev0
• Pytorch 2.3.0+cu121
• Datasets 2.19.2
• Tokenizers 0.19.1