How can I run it in hermes-agent with ollama
I see zero documents on how to run this model, meanwhile it is a full BP16 model, so it is another "it works in my machine and no one else" model?
Given the recent bug #136486 I see even its own product cannot run Qwen 3.8 27B, how the heck does this model release so soon, without its own tech support?
if you look at the page for the model, there's already 49 (!!) quantized versions of this model that will be happy to run on whatever hardware you have. At this point, day and age, nobody is going to write you specific instructions for you to load a random model in an outdated backend. if you don't understand how to do the basics, ask Claude or ChatGPT, even the free versions will be happy to help (or god forbid, do a web search)
The bug you referenced is specific to Hermes, which is a frontend, not a backend, or a model, and not directly tied to Qwen, or this finetune.
I think the OG model is nowhere "random", meanwhile backends are being updated daily, especially the "hermes frontend" (is it an AI harness?) still being rolling updated daily.
I have asked Claude and I am banned instantly haha, maybe I am going a bit too liberal.
Happy to look for a specific guide for this hot model, hopefully backed by its own product, instead of DIY slops.
Edit: Any ETA to the cloud API?
"How to run a quantized version of [this model] (link) on my machine that has [your VRAM] using ollama?" got you banned? I do not believe that.
Hermes is harness (or frontend, same thing), yes. That's the bug report you linked randomly in the middle of your post. it's about Hermes, not about the model.