Converting model not working due to Failed to detect model architecure

[whackl]

Hi there ,
when i try to convert a moel from huggingface with llama.cpp I get following error when i start the python conversion script :

kiuser@kisystem:/opt/huggingface/ollama-work/llama.cpp$ /opt/huggingface/ollama-work/bin/python3 convert_hf_to_gguf.py ../colqwen2-hf --outfile colqwen2-v1.0.gguf --outtype q8_0 INFO:hf-to-gguf:Loading model: colqwen2-hf Traceback (most recent call last): File "/opt/huggingface/ollama-work/llama.cpp/convert_hf_to_gguf.py", line 9485, in <module> main() File "/opt/huggingface/ollama-work/llama.cpp/convert_hf_to_gguf.py", line 9450, in main model_architecture = get_model_architecture(hparams, model_type) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/opt/huggingface/ollama-work/llama.cpp/convert_hf_to_gguf.py", line 9380, in get_model_architecture raise ValueError("Failed to detect model architecture") ValueError: Failed to detect model architecture
How to fix this ?

Regards ...

[misutoneko]

You might have better luck with the merged variant:
https://huggingface.co/vidore/colqwen2-v1.0-merged

If I've understood this correctly, the one you were using was just a LoRa extension (so maybe convert_lora_to_gguf.py would be more appropriate, if you wanted to do the merge yourself). But I've never used LoRas myself so idk.