DynamicTune – Closed-form trajectory weight surgery from 4B into 0.8B

The open-weight preview of Qwen4

Contribute to Mega4alik/ollm development by creating an account on GitHub.

The 397B native multimodal agent with 17B active params

I built slotstream, a way to run Qwen3.8-Flash-Next 4-bit on a low-memory mac starting from 16GB, a 125B parameter model that would need 100GB+ memory/RAM, thanks to expert-offloading/ssd-streaming. Easy to install/update, and mac-native using MLX and Swift. It ships with auto-mode, which makes a good tradeoff between memory usage and speed. I'll be implementing and porting the MTP module for speculative decoding next Local models really are the future of computing!

Features include: - Several preset periodic orbits: the classic Figure-8, plus newly discovered 3D solutions from Li and Liao's recent database of 10,000+ orbits (https://arxiv.org/html/2508.08568v1) - Full 3D camera controls (rotate/pan/zoom) with body-following mode - Force and velocity vector visualization - Timeline scrubbing to explore the full orbital period The 3D presets are particularly interesting. Try "O₂(1.2)" or "Piano O₆(0.6)" from the Load Presets menu to see configurations where bodies weave in and out of the orbital plane. Most browser…

Fine-tune LLMs from one YAML. Layer streaming trains an 8B model on a 4 GB laptop GPU. - MakazhanAlpamys/Soup