ner-test3

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: 1.1173
• Precision: 0.7826
• Recall: 0.8182
• F1: 0.8
• Accuracy: 0.7826

Model description

Fine-tuned Transformer based on the distilBERT architecture using Pytorch for detecting: Timestamps, KV and IPs.

Intended uses & limitations

Can be used on any system log containing timestamps, keyvalues and ips.

Training and evaluation data

Trained over 12000 logs: 3000 Apache, 1000 Csv, 1000 Dns, 3600 KV, 1000 Syslog and 3100 Miscellaneous logs. Evaluated on a small corpus of unseen logs labelled by hand.

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5e-05
• train_batch_size: 50
• eval_batch_size: 50
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 2

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
1.6299	1.0	1	1.2697	0.6522	0.6818	0.6667	0.6522
1.2767	2.0	2	1.1173	0.7826	0.8182	0.8	0.7826

Framework versions

• Transformers 4.29.2
• Pytorch 2.0.1
• Datasets 2.12.0
• Tokenizers 0.13.3