Add fraud-detection

README.md

@@ -16,17 +16,15 @@ You can install it with `pip install adaptive-classifier`.
 ## Model Details
 
 - Base Model: answerdotai/ModernBERT-base
-- Number of Classes: 4
-- Total Examples: 320
+- Number of Classes: 2
+- Total Examples: 161
 - Embedding Dimension: 768
 
 ## Class Distribution
 
 ```
-legitimate: 80 examples (25.0%)
-likely_fraud: 80 examples (25.0%)
-needs_investigation: 80 examples (25.0%)
-suspicious: 80 examples (25.0%)
+fraudulent: 80 examples (49.7%)
+legitimate: 81 examples (50.3%)
 ```
 
 ## Usage
@@ -50,7 +48,7 @@ classifier.add_examples(texts, labels)
 
 ## Training Details
 
-- Training Steps: 10
+- Training Steps: 6
 - Examples per Class: See distribution above
 - Prototype Memory: Active
 - Neural Adaptation: Active

config.json

@@ -32,17 +32,13 @@
   },
   "embedding_dim": 768,
   "id_to_label": {
-    "0": "suspicious",
-    "1": "needs_investigation",
-    "2": "legitimate",
-    "3": "likely_fraud"
+    "0": "legitimate",
+    "1": "fraudulent"
   },
   "label_to_id": {
-    "legitimate": 2,
-    "likely_fraud": 3,
-    "needs_investigation": 1,
-    "suspicious": 0
+    "fraudulent": 1,
+    "legitimate": 0
   },
   "model_name": "answerdotai/ModernBERT-base",
-  "train_steps": 10
+  "train_steps": 6
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:dd67e0a6638ae494588e7e664453913a55c06c6376800f753e5b157edfd21b41
-size 3562920
+oid sha256:65773e1ee4b2d3bc1c047b5b15431ec024c6beaab697a092e31584310e56f7a1
+size 3553512