cynthiachan/finetune-roberta

<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->

training

This model is a fine-tuned version of roberta-base on the cynthiachan/FeedRef2022 dataset. It achieves the following results on the evaluation set:

• —Loss: 0.0342
• —Attackid Precision: 0.9444
• —Attackid Recall: 0.9808
• —Attackid F1: 0.9623
• —Attackid Number: 52
• —Bitcoinaddr Precision: 0.7778
• —Bitcoinaddr Recall: 1.0
• —Bitcoinaddr F1: 0.8750
• —Bitcoinaddr Number: 7
• —Cve Precision: 0.9118
• —Cve Recall: 0.9841
• —Cve F1: 0.9466
• —Cve Number: 63
• —Defenderthreat Precision: 1.0
• —Defenderthreat Recall: 1.0
• —Defenderthreat F1: 1.0
• —Defenderthreat Number: 7
• —Domain Precision: 0.9429
• —Domain Recall: 0.9612
• —Domain F1: 0.9519
• —Domain Number: 206
• —Email Precision: 0.875
• —Email Recall: 0.9767
• —Email F1: 0.9231
• —Email Number: 43
• —Filepath Precision: 0.9552
• —Filepath Recall: 0.9667
• —Filepath F1: 0.9609
• —Filepath Number: 1654
• —Fingerprint Precision: 0.0
• —Fingerprint Recall: 0.0
• —Fingerprint F1: 0.0
• —Fingerprint Number: 2
• —Hostname Precision: 0.8861
• —Hostname Recall: 0.9722
• —Hostname F1: 0.9272
• —Hostname Number: 144
• —Ipv4 Precision: 0.9828
• —Ipv4 Recall: 1.0
• —Ipv4 F1: 0.9913
• —Ipv4 Number: 171
• —Ipv6 Precision: 0.6
• —Ipv6 Recall: 1.0
• —Ipv6 F1: 0.7500
• —Ipv6 Number: 3
• —Md5 Precision: 0.9767
• —Md5 Recall: 0.9952
• —Md5 F1: 0.9859
• —Md5 Number: 421
• —Sha1 Precision: 0.9778
• —Sha1 Recall: 0.8980
• —Sha1 F1: 0.9362
• —Sha1 Number: 49
• —Sha256 Precision: 0.9440
• —Sha256 Recall: 0.9959
• —Sha256 F1: 0.9693
• —Sha256 Number: 491
• —Uri Precision: 1.0
• —Uri Recall: 0.6364
• —Uri F1: 0.7778
• —Uri Number: 11
• —Overall Precision: 0.9510
• —Overall Recall: 0.9744
• —Overall F1: 0.9626
• —Overall Accuracy: 0.9917

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
• —trainbatchsize: 8
• —evalbatchsize: 8
• —seed: 42
• —optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• —lrschedulertype: linear
• —num_epochs: 3.0

Training results

Framework versions

• —Transformers 4.24.0
• —Pytorch 1.12.1+cu113
• —Datasets 2.6.1
• —Tokenizers 0.13.2