Error when trying to convert a HF model which is a LORA PEFT fine tuned version of phi-128k

[swarnava112]

I am trying to convert a HF model I finetuned, the model is an adapter model for microsoft phi-128k. I donot know what arch should I specify as I didnot see "phi" as an arch option, when I looked into the code.

Any help is appriciated.

[arnfaldur]

llama.cpp does not support phi-3-128k currently #6849
The work on adding it is happening here #7225

Please try searching a little before opening an issue. Feel free to close it as well as it's already being tracked.

[ngxson]

The convert-lora-to-ggml script has been deprecated and already been removed: #7204