Dev
GitHub repos gaining traction - what high-signal users are starring and what's climbing the board, captured daily and enriched from GitHub. Raw material for spotting new tech and patterns worth building on.
72
repos tracked
72
surfaced this week
25
created < 30d
Python
top language
72 repos
-
5KB WebGL globe lib.
-
KVarN is a native vLLM KV-cache quantization backend for your agents: 3-5x more context, throughput above FP16, and FP16-level accuracy. Calibration-free, one flag.
-
Fast LLM inference with Elixir and Bumblebee
-
An open, technical and introductory book for the Zig programming language 📚📖
-
ConTrack: Constrained Hand Motion Tracking with Adaptive Trade-off Control
-
An interactive RLM (Recursive Language Model) agent built on DSPy.
-
Open-source infrastructure for Computer-Use Agents. Sandboxes, SDKs, and benchmarks to train and evaluate AI agents that can control full desktops (macOS, Linux, Windows).
-
Continual Learning Bench
-
Official repository for Parallax (Parameterized Local Linear Attention)
-
A Claude Skill to create an evidence review to inform specific health questions
-
SDPG: Self-Distilled Policy Gradient
-
GDM Science Skills to speed up agentic scientific workflows with better grounding and higher token efficiency. Integrate insights from AlphaGenome, AFDB, UniProt and 30+ other databases and tools.
-
Skills for the Gemma and model/agent interactions
-
Experimental fork of Karpathy's nanoGPT modifying the architecture and training loop to make per-source influence over predictions identifiable and controllable.
-
Project the aircraft passing overhead onto your ceiling in real time, from an RTL-SDR — with a live sky layer (sun, moon, stars, ISS) and where each plane is headed.
-
high-performance inference and serving library for interactive autoregressive video and world models
-
Ideogram 4: Open image model at the forefront of design
-
Miso TTS is an 8 billion, highly emotive text-to-speech model
-
Use Garry Tan's exact Claude Code setup: 23 opinionated tools that serve as CEO, Designer, Eng Manager, Release Manager, Doc Engineer, and QA
-
Desktop version of Pytorch memory viz, handling profiles of 10GB and more, with quality-of-life improvements
-
d4 Model Counter
-
Data Science Skills for AI agents like Claude Code
-
Deploy any VLA model to any edge GPU with one command. Verified cos=+1.0 ONNX parity. Open source.
-
torch_remat fine-grained activation checkpointing API
-
Keep it Simple, Stupid AI Agent Framework with a General-Purpose and Software Engineering Assistant, called KISS Sorcar.
-
Wuji Hand in-hand reorientation RL with sim-to-real deployment, built on mjlab
-
NVIDIA Cosmos is an open platform of world models, datasets, and tools that enables developers to build Physical AI for robots, autonomous vehicles, smart infrastructure, and more.
-
A minimal codebase to generate synthetic coding agent session traces
-
A unified toolkit for detecting AI-generated content
-
🚀 An open-source, hands-on curriculum bridging the gap from basic RL concepts to LLM alignment, RLVR, and advanced Agentic systems.
-
★ 12.3k zakirullin/cognitive-load🧠 Cognitive load is what matters
-
A fast, helpful, and open-source document parser
-
Parakeet implementation in C++ with ggml
-
Code for "Flat-Pack Bench: Evaluating Spatio-Temporal Understanding in Large Vision-Language Models through Furniture Assembly" (CVPR 2026)
-
XoRL
-
Finds potential accessibility gaps, files GitHub issues to track them, and attempts to fix them with Copilot.
-
Post-training framework for large models, from new objectives to new rollout systems.
-
A curated collection of papers and resources on On-Policy Distillation for Large Language Models.
-
Use agent to learn agent - A skeleton course on how to design, build, and operate production AI agents
-
JAX in JavaScript – ML library for the web, running on WebGPU & Wasm
-
⚡ TabPFN: Foundation Model for Tabular Data ⚡
-
Add secure sharing, document tracking, and workflow automation on top of your storage (Nextcloud, Google Drive, Dropbox). Self-hosted DocSend alternative.
-
Miles is an enterprise-facing reinforcement learning framework for LLM and VLM post-training, forked from and co-evolving with slime.
-
An extensive and commented list of resources on Late-Interaction Multivector Retrieval.