[Errno 2] No such file or directory: './llama.cpp/llama-quantize'

#140
by AlirezaF138 - opened

I was using the space around 12 hours ago, and everything was fine, but now I get this error:
[Errno 2] No such file or directory: './llama.cpp/llama-quantize'

Same it just doesn't seem to like some models, also had trouble using llama.cpp for quants on CoLab- I wish I knew what it was (Honestly, I have gotten some models that don't work on gguf-my-repo others no problem but I'm a bit new and figured it was me.

I have the same problem. Really sad because I'm in the middle of testing out new merges and I can't use the space to quantize in Q6_K. Also, this error might not be related to specific models but maybe the source file? I tested it with my old model, which I already successfully quantized before using this, but now it also won't go through anymore. So maybe the problem doesn't lie with the models themselves. Maybe something happened with the llama-quantize being mentioned?

ggml.ai org

This error is due to Makefile being deprecated recently in llama.cpp. After migrating to cmake, it's now working normally

ngxson changed discussion status to closed

just tested now , still same error

Confirmed just now, the space still produces the same error.

The thread was marked closed, but the problem remains. Can this thread still notify the authors, or do we have to create a new thread with the same error report?

Capture.JPG

ggml.ai org

Ok seems like another error, I'll have a look

ngxson changed discussion status to open

Any updates here?

ggml.ai org

sorry for the delay, the cmake build failed due to -DLLAMA_CURL. I've just removed it, the space is rebuilding now

ggml.ai org

It should work now

image.png

Thank you so much! Confirmed working now.

image.png

ngxson changed discussion status to closed

Sign up or log in to comment