Downloads · 30 days
0
nspam
barrydeen/nspam
nspam is a text classification model from barrydeen. Use it when you need a label for a piece of text. It is set up for lightgbm. The card lists the license as mit.
Scores bundles of Nostr kind:1 reply notes from a single author as real or bot. Designed for mobile clients (Kotlin / Swift / JS).
Access
Public
Updated Aug 30, 2026
Repo size
48.3 MB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.json47.4 MB · 72%
At a glance
- Task
- Text Classification
- Library
- lightgbm
- License
- mit
- Access
- Public
- Created
- Apr 15, 2026
- Updated
- Aug 30, 2026
- SHA
- 29f76c80
Model card
From the Hugging Face model README
nspam — on-device Nostr reply-spam classifier
Scores bundles of Nostr kind:1 reply notes from a single author as
real or bot. Designed for mobile clients (Kotlin / Swift / JS).
- Model type: LightGBM gradient boosted trees over hashed character + word n-grams and structural features.
- Input: 1–10 recent reply notes (events with
etags) from one pubkey. - Output: calibrated probability ∈ [0, 1] that the author is a reply-spammer.
- Runtime: LightGBM4j (JNI) or ONNX Runtime on device.
Intended use
Client-side filtering/ranking of replies in Nostr apps. Use the score to deprioritize or hide likely spam replies. Combine with user mutes and follow graph rather than hard-blocking on a single score.
Holdout metrics (v2.4)
| metric | value |
|---|---|
| average precision | 0.9567 |
| ROC AUC | 0.9721 |
| precision @ recall 0.9 | 0.9439 |
| per-author accuracy | 0.9489 |
Limitations
- Reply-focused. Trained on reply notes only — not designed for scoring feed posts or other event kinds.
- English-heavy training data. Non-Latin scripts underrepresented.
- Adversarial drift. Spammers evolve. Retrain periodically.
- Cold start. Accounts with <3 replies have limited signal.
License
MIT.
- Task
- Text Classification
- Library
- lightgbm
- License
- mit
- Languages
- multilingual
- Created
- Apr 15, 2026
- Updated
- Aug 30, 2026