AttributeError: function 'llama_sampler_init_tail_free' not found after compiling llama.pcc with hipBLAS
Opened this issue · 2 comments
Micromanner commented
Prerequisites
- [ x] I am running the latest code. Development is very rapid so there are no tagged versions as of now.
- [ x] I carefully followed the README.md.
- [ x] I searched using keywords relevant to my issue to make sure that I am creating a new issue that is not already open (or closed).
- [x ] I reviewed the Discussions, and have a new bug or useful enhancement to share.
Expected Behavior
Run the model
Current Behavior
I get an attribute error when importing the module.
# Environment and Context
Windows 10, 7900XTX, 5800X3D, python3.10.11, gcc version 13.2.0
# Failure Information (for bugs)
I think its a bug, I have no problem running llama.cpp, only the python package is making problems.
# Steps to Reproduce
1. Compile llama.cpp for hipBLAS (I followed the instructions)
2. python -m pip install . --no-cache-dir
3. Run my code (model) with python run.py
Failure Logs
Traceback (most recent call last):
File "D:\test\run.py", line 7, in <module>
from llama_cpp import Llama
File "D:\test\env\lib\site-packages\llama_cpp\__init__.py", line 1, in <module>
from .llama_cpp import *
File "D:\test\env\lib\site-packages\llama_cpp\llama_cpp.py", line 3201, in <module>
def llama_sampler_init_tail_free(z: float, min_keep: int) -> llama_sampler_p:
File "D:\test\env\lib\site-packages\llama_cpp\_ctypes_extensions.py", line 113, in decorator
func = getattr(lib, name)
File "C:\Users\01tee\AppData\Local\Programs\Python\Python310\lib\ctypes\__init__.py", line 387, in __getattr__
func = self.__getitem__(name)
File "C:\Users\01tee\AppData\Local\Programs\Python\Python310\lib\ctypes\__init__.py", line 392, in __getitem__
func = self._FuncPtr((name_or_ordinal, self))
AttributeError: function 'llama_sampler_init_tail_free' not found. Did you mean: 'llama_sampler_init_min_p'?
jianlins commented
similar error when try to load https://huggingface.co/Qwen/Qwen2.5-Coder-7B-Instruct-GGUF
djeffkanda commented
I had the same issue, but solved it by pulling and installing the latest version. The fix is in the following commit[commit]