Update README.md

README.md

@@ -1,3 +1,78 @@
----
-license: apache-2.0
----
+---
+tags:
+- text-classification
+- security
+- blue-team
+- roberta
+license: apache-2.0
+datasets:
+- trendmicro-ailab/Primus-FineWeb
+metrics:
+- precision
+- recall
+- f1
+pipeline_tag: text-classification
+library_name: transformers
+---
+
+# BlueSecureBERT 🟦🛡️
+
+Detects **blue-team / defensive security** text (English).
+
+| Split       | Precision | Recall | F1    | F₂    | CE-loss | Threshold |
+|-------------|-----------|--------|-------|-------|---------|-----------|
+| Validation  | **0.949** | **0.991** | **0.969** | **0.982** | **0.011** | **0.579** |
+
+> **Recommended cut-off:** `prob >= 0.50` (arg-max on the validation split)
+
+---
+
+## Intended uses & limits
+
+* **Triage** incident reports, chat logs, or bug-bounty write-ups  
+* **Input language:** English  
+* **No external test set** yet → treat numbers as optimistic
+
+---
+
+## Training data
+
+| Label     | Rows    |
+|-----------|---------|
+| Offensive | 30 746  |
+| Defensive | 19 550  |
+| Other     | 130 000 |
+| **Total** | **180 296** |
+
+Source: [Primus-FineWeb](https://huggingface.co/datasets/trendmicro-ailab/Primus-FineWeb)
+
+---
+
+## Model details
+
+| Field          | Value                                                |
+|----------------|------------------------------------------------------|
+| Base encoder   | `ehsanaghaei/SecureBERT` (RoBERTa-base, 125 M)       |
+| Objective      | One-vs-rest, focal-loss (γ = 2)                       |
+| Training       | 3 epochs · micro-batch 16 · LR 2e-5                   |
+| Hardware       | 1× RTX 4090 (≈ 41 min)                               |
+| Inference dtype| FP16-safe                                            |
+
+---
+
+## Quick start
+
+```python
+from transformers import pipeline, AutoModelForSequenceClassification, AutoTokenizer
+
+model_id = "HagalazAI/BlueSecureBERT"
+tok      = AutoTokenizer.from_pretrained(model_id)
+model    = AutoModelForSequenceClassification.from_pretrained(model_id)
+clf      = pipeline("text-classification", model=model, tokenizer=tok, top_k=None)
+
+text = "Investigate potential SQL injection vulnerabilities."
+prob = clf(text)[0]["score"]  # sigmoid prob for class 0 (Defensive)
+print(f"P(defensive) = {prob:.3f}")
+
+is_blue = prob >=  0.579        # ← recommended threshold
+print("is_blue:", is_blue)