yfyeung/prunedctc
Memory-efficient CTC loss with exact loss and first-order gradient equivalence under vocabulary reduction, and activation memory that no longer scales linearly with vocabulary size.
01 / MACHINE INTELLIGENCE
Models you can run. Agents you can inspect. Tools you can build on.
← All topicsMemory-efficient CTC loss with exact loss and first-order gradient equivalence under vocabulary reduction, and activation memory that no longer scales linearly with vocabulary size.
Open the project to explore the idea and the code.
Claude Code mod: hand a long session off to a fresh one with a structured brief, instead of auto-compacting
Open the project to explore the idea and the code.
Open-source hotel & short-term rental PMS - front desk, booking engine, folios & GST/tax billing, housekeeping, POS, WhatsApp. Agent-ready via MCP (Claude). Built on Frappe. AGPL-3.0.
Typed-decision models (noul / choice / score) trained by a self-improving loop of AI agents — checkpoints, the code that produced them, and every version that failed.
Claude Fable 5.1 prompting guide as Claude Code skills (KO/EN/ZH). Output quality goes up on Opus 5 and Sonnet 5 too.
OmniEdu: Open Foundation Models for Learning and Teaching
A sovereign, zero-trust local AI agent and micro-IDE.
A local-first workspace for persistent AI agents. Built with Bun, React, SQLite, and Pi.
[NeurIPS 2026 Spotlight] Official code and checkpoints for OneCanvas: 3D Scene Understanding via Panoramic Reprojection.
Agent skills for daily use.
Coworker is a simple desktop AI assistant that can search the web, create files, schedule tasks, use memory, and run with local or cloud AI models.
🎮 300 source-verified AI-built games — rated, categorized, and linked to real GitHub code.
Claude Code skill: hand bulk judgment calls to Jev. 86% fewer Claude tokens on a 12-task benchmark, up to 20x faster. One-line npx install.
Agentic AI security scanner that hunts exploitable vulnerabilities like an adversary, proves them with executable PoCs, and fixes them test-first. A maintained fork of Capital One's VulnHunter, rebuilt for any agent harness.
vLLM inference solutions for NVIDIA GB10 (DGX Spark, sm121) — KIMI-K3 B12X_MLA + DSpark, GLM-5.2
4DCodeBench: Benchmarking Agents on Inverse Graphics of Dynamic Scenes
A high-performance, GGUF-native Rust & CUDA inference engine optimized for cold-start latency and real-time 'System 1' agent decision loops.
Official PyTorch implementation of VisionHOPE: Visual Backbones as Self-Modifying Learning Systems.
Evidence-based code review skills for Codex and Claude Code
An unofficial local visualizer for Jev (TypeSafe): a live view of every call your code makes. Not affiliated with TypeSafe AI.
Deterministic security for real-world agentic applications
Open-source skills that empower any AI agent (Claude, Cursor, Hermes, etc.) to generate end-to-end marketing campaigns - UGC videos, ad videos, product photography and more.