vision-language models on consumer gear with llama.cpp

[rocmhouse90]

What worked for me was pretty simple. If you're on MLX, I'd confirm the version first and then reduce variables until the behavior is obvious.

Interesting, I hadn't tried that. I don't think your conclusion is far off. Worth rechecking your defaults though, because one bad assumption can make the rest of the thread look confusing.

[cachelab]

I'd approach it in a boring order. Before rebuilding anything, I'd test one clean baseline with Mixtral and compare from there.

[visionforge]

My first guess would be this: Start by verifying the driver/runtime combo, then test with a smaller prompt, then look at NVFP4. Once you know the plain path works, then add the faster path back in piece by piece.