Newbie trying to convert llama models to gguf

[malavi3]

Hi everyone,

Sorry for the noob question in advance.

Basically I have a llama 3.2 model from HF that I'm trying to convert to GGUF but I get the below error:

File "/home/me/llamacpp/build/bin/convert_hf_to_gguf.py", line 2580, in <module>
    class PhiMoeModel(Phi3MiniModel):
  File "/home/me/llamacpp/build/bin/convert_hf_to_gguf.py", line 2581, in PhiMoeModel
    model_arch = gguf.MODEL_ARCH.PHIMOE
                 ^^^^^^^^^^^^^^^^^^^^^^
AttributeError: type object 'MODEL_ARCH' has no attribute 'PHIMOE'

The command I'm running to convert my model is below:

python3 convert_hf_to_gguf.py --model "my/model/path" --outtype f16

Anyone has any clue what could cause the error?

[malavi3]

I found the issue. In case any other new ppl go through the same:

I had downloaded the llama.cpp zip for ubuntu from the release page + downloaded the convert_hf_to_gguf.py separately.

Looks like the zip doesn't include the dependencies for convert_hf_to_gguf.py.

So I went through cloning the repo and using cmake to install llama.cpp fully and it fixed the issue.