Instructions to use sentinelsage/arbiter-log-classifier with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use sentinelsage/arbiter-log-classifier with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-classification", model="sentinelsage/arbiter-log-classifier")# pip install -U transformers accelerate # Load model directly from transformers import AutoTokenizer, AutoModelForSequenceClassification tokenizer = AutoTokenizer.from_pretrained("sentinelsage/arbiter-log-classifier") model = AutoModelForSequenceClassification.from_pretrained("sentinelsage/arbiter-log-classifier", device_map="auto") - Notebooks
- Google Colab
- Kaggle
Add model card for Oracle Shield
Browse files
README.md
CHANGED
|
@@ -1,3 +1,62 @@
|
|
| 1 |
---
|
| 2 |
license: apache-2.0
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
license: apache-2.0
|
| 3 |
+
library_name: transformers
|
| 4 |
+
tags:
|
| 5 |
+
- distilbert
|
| 6 |
+
- text-classification
|
| 7 |
+
- security
|
| 8 |
+
- oracle-shield
|
| 9 |
+
pipeline_tag: text-classification
|
| 10 |
---
|
| 11 |
+
|
| 12 |
+
# Oracle Shield — Log threat classifier
|
| 13 |
+
|
| 14 |
+
## What this is
|
| 15 |
+
|
| 16 |
+
A **DistilBERT** sequence-classification checkpoint fine-tuned for Oracle Shield.
|
| 17 |
+
|
| 18 |
+
Given a short security-log style string, it predicts one of four actions:
|
| 19 |
+
|
| 20 |
+
| Index | Action (app mapping) |
|
| 21 |
+
|------:|----------------------|
|
| 22 |
+
| 0 | block |
|
| 23 |
+
| 1 | quarantine |
|
| 24 |
+
| 2 | warn |
|
| 25 |
+
| 3 | none |
|
| 26 |
+
|
| 27 |
+
The Oracle Shield app then scales a threat score from the predicted class / probabilities and may adjust the final recommendation with a separate rule-based layer.
|
| 28 |
+
|
| 29 |
+
## Project
|
| 30 |
+
|
| 31 |
+
Part of **Oracle Shield** (Django + React): paste a security log or ask a cybersecurity question.
|
| 32 |
+
|
| 33 |
+
This repository holds **weights only**. Application code is in the public Oracle Shield GitHub repository.
|
| 34 |
+
|
| 35 |
+
## Training note
|
| 36 |
+
|
| 37 |
+
This is a **custom fine-tune**, not an unmodified public DistilBERT base checkpoint. Do not treat Hub `distilbert-base-uncased` (or similar) as a drop-in substitute for this artifact.
|
| 38 |
+
|
| 39 |
+
## Intended use
|
| 40 |
+
|
| 41 |
+
- Running the Oracle Shield full-ML path locally
|
| 42 |
+
- Experimentation and educational use with the companion application
|
| 43 |
+
|
| 44 |
+
**Not** a production SOC product. No warranty. Outputs can be wrong; do not use for live blocking, compliance, or incident response without your own validation.
|
| 45 |
+
|
| 46 |
+
## How Oracle Shield loads it
|
| 47 |
+
|
| 48 |
+
Local path after download:
|
| 49 |
+
|
| 50 |
+
```text
|
| 51 |
+
backend/models/log_based_model/
|
| 52 |
+
```
|
| 53 |
+
|
| 54 |
+
Loaded with Hugging Face `DistilBertForSequenceClassification.from_pretrained(<local_or_hub_id>)` when `ORACLE_DEMO_MODE=false`.
|
| 55 |
+
|
| 56 |
+
## Files
|
| 57 |
+
|
| 58 |
+
Expect a standard Transformers export, e.g. `model.safetensors`, `config.json`, tokenizer files (`vocab.txt`, `tokenizer_config.json`, …).
|
| 59 |
+
|
| 60 |
+
## License
|
| 61 |
+
|
| 62 |
+
Apache License 2.0. Demo only — provided as-is, without warranty of any kind.
|