Prince Canuma
Prince Canuma
> Found the issue! > This version points to llava in the model config. I patched it locally. > Don't worry, I will add a condition to fix this at...
Could you help me with that ? Also, perhaps adding a way to scan for models on the mlx-community based on names ?
Closing stale
I also think it will probably fix the broadcast error reported [here](https://github.com/Blaizzy/mlx-vlm/issues/194). I tried adding print statements to idefics3.swift so I can see the tensors but I can't see them...
> @Blaizzy Unfortunately this PR didn't solve that issue. It crash [here](https://github.com/ml-explore/mlx-swift-examples/blob/7109e3ee7b34776efc78990b564612bef7b0c27e/Libraries/MLXVLM/Models/Idefics3.swift#L553) before entering prepareInputsForMultimodal(). > > Actual crash happens on [ops.cpp](https://github.com/ml-explore/mlx-c/blob/e8889ddf56b3bd734eb79d5e5b13586a39c01403/mlx/c/ops.cpp#L28) but since this is shape bug, I think...
> As far as I can tell we are passing a non-split image in where a split image is expected. I don't know if that is the entirety of it,...
@davidkoski thanks a lot for the tips! > I typically use the debugger in Xcode for this. You can press cmd-opt-r to set the arguments to the program. You may...
Thanks! I'm running MacOS :)
@benzimring that's perfect. Please do! I have been a bit busy with v0.1.0 and work but I'm back.
Could you send a PR with your proposed changes?