I got a lot of random empty space in the python script generated

[BVEsun]

llama.cpp version: llama-b4938-bin-win-cuda-cu12.4-x64.zip
llm: bartowski/mistralai_Mistral-Small-3.1-24B-Instruct-2503-GGUF/mistralai_Mistral-Small-3.1-24B-Instruct-2503-Q4_K_M.gguf

llama-server.exe -m ..\models\mistralai_Mistral-Small-3.1-24B-Instruct-2503-Q4_K_M.gguf -ngl 99 --temp 0.6 -fa -c 20480 --metrics

I ask it to rewrite a python script,
the script is looking fine,
but it insert a lot of random empty space in the python script which make it not working,

Do anyone have similiar issue or just mine?

[ejrydhfs]

I don't use that model, however this usually happens due to limitations of the model itself. Also, quantization can impair code generation or so I have heard.