Openclaw to ollama is slow, model is fast
When I use openclaw and back an agent with ollama on an m4-pro, it constantly hangs, or times out.
When I curl the model directly from openclaw's workspace in a k8s pod, it's insanely fast. We're talking 66 tokens/s with a 3.9s response time from curl to end with
How can I track down where the bottleneck is? It doesn't appear to be the underlying infra or the backend model. Is it my openclaw config?
When I curl the model directly from openclaw's workspace in a k8s pod, it's insanely fast. We're talking 66 tokens/s with a 3.9s response time from curl to end with
stream=trueHow can I track down where the bottleneck is? It doesn't appear to be the underlying infra or the backend model. Is it my openclaw config?
The AI that actually does things. Emails, calendar, home automation - all from your favorite chat app. New shell, ready to help. The lobster way.
130,529Members
Resources
Recent Announcements
Similar Threads
Was this page helpful?
