@NousResearch This is super cool! It would be lovely if y'all cleaned up the dependencies you've taken on deprecated modules. There are some gnarly warnings in there:
256K context and multimodal (text and vision) in 16GB of VRAM?! I’m so gonna try this… I taught Hermes how to swap models on the fly (so I could bounce between llama.cpp hosted Qwen text model and a ComfyUI hosted vision model for image gen… but I can’t validate the results. 🤞this solves some of my workflow issues!
@grok@Trameroo@k1rallik Thank you. I missed that because the prospectus doesn’t call it out — it’s a platform-wide Fidelity rule to discourage flipping.
@WickedAIyoutube@Ducnghia16 Instant classic. The lyrics are amazing... how much of them were written by AI vs by hand? They seem to good to be pure AI!
@sudoingX Awesome! 👏. I’ve got a 3090 FE (24GB).
Machine has 128GB + 13900K. Usually run Windows 11 for Workstations, but I’m considering Linux.
What Linux distro would you recommend for the smoothest Hermes operation?
I assume native Linux is preferred to WSL or Docker on Win?