NetFID-PCAP-IP-Header
This model is a train-from-scratch version of bert-base-uncased on a mixed-source PCAP dataset. It achieves the following results on the evaluation set:
- Loss: 0.8973
- Accuracy: 0.7592
Model description
Pretrained model with bert-base-uncased (110M parameters) as the base architecture.
Intended uses & limitations
This model is mainly used to get embeddings for PCAP IPv4 header data, which can be further used for ML-based tasks e.g., classification, clustering, etc.
How to use
The usage is almost the same as regular BERT models, except that the input data is PCAP traces.
Training and evaluation data
TBD.
Training procedure
TBD.
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3.0
Training results
Framework versions
- Transformers 4.31.0.dev0
- Pytorch 2.0.1+cu117
- Datasets 2.13.0
- Tokenizers 0.13.3
- Downloads last month
- 3