# Laya spam classifiers decision model by Nikita Kolmogorov > Laya spam classifiers copy Jev judgments for Telegram and Farcaster. Public weights include reported teacher agreement and task-specific thresholds. Updated 2026-11-07 (Australia/Adelaide) ## Quick answers ### What is Laya spam classifiers by Nikita Kolmogorov? [Laya spam classifiers](): Two multilingual [Laya]() fine-tunes copy [Jev]() judgments for Telegram messages and Farcaster casts or accounts. [Laya spam classifiers facts]() [Source]() ### Who makes Laya spam classifiers? [Laya spam classifiers](): Nikita Kolmogorov [Laya spam classifiers facts]() [Source]() ### Are Laya spam classifiers's weights open? [Laya spam classifiers](): Public weights, question files and loading examples for both classifiers Licence: Apache-2.0, according to both model cards [Laya spam classifiers facts]() [Source]() By Nikita Kolmogorov · Hugging Face package release: 2026-11-06 · OPEN Two multilingual [Laya]() fine-tunes copy [Jev]() judgments for Telegram messages and Farcaster casts or accounts. The Telegram bot uses a 0.81 deletion gate and exempts admins. Farcaster ships Platt scaling for its raw scores. The evaluations measure agreement with the teacher, including its mistakes. ## Facts - Base: convaiinnovations/laya-multilingual - Inputs: Text, links and metadata; no media-content or avatar inspection - Training: Telegram: 5,540 messages, with training data unreleased; Farcaster: 16.6k casts and 4,384 account snapshots - Licence: Apache-2.0, according to both model cards - Access: Public weights, question files and loading examples for both classifiers ## Claims and sources - **VENDOR-REPORTED**: The maker reports 95.4% agreement with [Jev]() on 978 held-out Telegram messages at the 0.81 gate, precision 0.82 and recall 0.70. These compare with [Jev]()'s delete labels, not human judgments. Reported median latency is 38 ms on an Apple GPU and about 400 ms on a few CPU threads. [source]() - **VENDOR-REPORTED**: The maker reports 95.7% agreement on 2,489 held-out Farcaster casts at gate 0.5, precision 0.92 and recall 0.80. Account agreement is 97.1% on 652 held-out accounts. The newest 15% is held out by time; new spam styles may differ. [source]() ## Links - [Source: Telegram card and weights]() - [Source: Farcaster card and weights]() ## Related projects No related projects are listed in the registry.