YogeshTembe

@dosu-bot I am trying to use "Ctransformers" library from langchain and you are referring to "HuggingFacePipeline" library. Can you please suggest some way to use gpu with "Ctransformers" library in...

@dosu-bot Yes sure. Here is the required information - 1) langchain version = 0.0.306 , ctransformers version = 0.2.27 2) CUDA version = 12.2 3) There is no error or...

@dosu-bot The solution provided does not work. Also Accelarate is not part of langchain as well. `from langchain.llms import Accelerate` does not work.

@mzen17 @radames I tried following commands on windows but gpu is not utilised. CMAKE_ARGS="-DLLAMA_CUBLAS=on" FORCE_CMAKE=1 pip install llama-cpp-python 1) set CMAKE_ARGS="-DLLAMA_CUBLAS=on" 2) set FORCE_CMAKE=1 3) pip install --upgrade --force-reinstall llama-cpp-python...

@radames Yes I have followed the same. We just need to set one variable right ? => CMAKE_ARGS = "-DLLAMA_OPENBLAS=on"