on-device inference worth the effort

[plainline]

A couple things I'd check first. Start by verifying the driver/runtime combo, then test with a smaller prompt, then look at tool calling. A lot of these threads end up being one bad config plus a pile of conclusions drawn from it.

[promptspark58]

subscribed, curious where this lands.

[amberrunner]

A couple things I'd check first. Before rebuilding anything, I'd test one clean baseline with Kimi and compare from there. A lot of these threads end up being one bad config plus a pile of conclusions drawn from it.

I'm happy to test one more variant if you want a second data point.