A library that leverages large language models (llms) for extracting topics from text documents.
To install for CPU run pip install llmtopic
.
To install with GPU support run CMAKE_ARGS="-DLLAMA_CUBLAS=on" pip install llmtopic
.
If you are under Ubuntu and encounter errors regarding a failed build consider upgrading your cuda version to >=12.1. Also, it could be you have to point to the new nvcc version but modifying the install command as follows: PATH="/usr/local/cuda-12.3/bin:$PATH" CMAKE_ARGS="-DLLAMA_CUBLAS=on" pip install llmtopic
. Note that you of course have to adjsut that to your specific CUDA version.
For vision application use this install command:
PATH="/usr/local/cuda-12.3/bin:$PATH" CMAKE_ARGS="-DLLAMA_CUBLAS=on -DLLAVA_BUILD=on" pip install --upgrade --force-reinstall --no-c ache-dir llama-cpp-python==0.2.55