Megathread for listing latest open source projects, research papers that are helping optimizations, efficiencies and accessibility to Open Source LLM and related hardware, software ?

Excerpt

I start with some informations gathered thorough endless posts reading on this sub and online: Inference and hardware optimization projects https://dwarfstar.sh/ - inference engine optimization proposal https://github.com/JustVugg/colibri - Treats VRAM + RAM + storage as one managed inference memory hierarchy, with expert streaming, caches and a strong focus on MoE. https://openfreedom.it/ - agentic harness proposal https://github.com/exo-explore/exo - Core source for heterogeneous/topology-awar