LM Studio help

[martymcfly1]

I've fine tuned a model successfully, tested it using Mlx-lm from the terminal and it works. I've tried to copy the full folder with the fused model to LM Studio and it loads without issue but giving it the same prompt as used in the testing phase in the terminal outputs the wrong answer - as if the fine tuning never took place. I don't know if it's an issue with Mlx-lm or LM studio. All I'm trying to do is fine tune a model to use in LM Studio. Can anybody please help me?

[awni]

Was it a LoRA fine tune? If so, did you fuse the adapters or not?

what’s the command you used to fine tune?

[martymcfly1]

Thanks for replying awn. It was a LoRA fine tune - I was following this example - https://github.com/filipw/2025-zurich-azure-ug-demos/tree/main/fine-tuning - The example checks out on my Mac, but if I copy the model folder created after the fuse command over to LM Studio on my Mac and try the same user prompts, I get no answer back at all, or a gibberish answer back.

[awni]

but if I copy the model folder created after the fuse

Could you try the model after fuse but outside of LM Studio? Possibly the fusing is the issue.

[awni]

Possibly related: #203 (comment)