I haven't managed to get it to work properly via ROCm on Strix Halo. With a lot of trial and error and AI support, the pipeline was running kind of stable, but the response time - even using tiny llm models - was not usable. I'm no expert though and I don't know if it is likely that this will be optimized for AMD hardware with Nvidia taking over hugging face. Anyway, I'd really appreatiate.
RK
Adamspliff
AI & ML interests
None yet
Recent Activity
commentedon an article about 1 month ago
Reachy Mini goes fully local new activity 4 months ago
pollen-robotics/reachy_mini_conversation_app:Plans for an official local LLM-backed version?Organizations
None yet
commented on Reachy Mini goes fully local about 1 month ago
Plans for an official local LLM-backed version?
7
#16 opened 6 months ago
by 
lesutton