[Question] Fail to pip install mlc-llm · Issue #2974 · mlc-ai/mlc-llm · GitHub
❓ General Questions
Hi, I’m trying to install mlc-llm on my Jetson agx orin.
 Environment: Jetson agx orin、Ubuntu 20.04、CUDA 12.2
 I ran the following commands:conda create -n mlc python=3.11conda activate mlcconda install -c conda-forge git-lfsconda install -c conda-forge libgcc-ng
 and thenpython -m pip install --pre -U -f https://mlc.ai/wheels mlc-llm-nightly-cu122 mlc-ai-nightly-cu122
 It seems that I have successfully installed them.Installing collected packages: mlc-llm-nightly-cu122, mlc-ai-nightly-cu122 Successfully installed mlc-ai-nightly-cu122-0.1 mlc-llm-nightly-cu122-0.1
 But I failed to import mlc_llmpython -c "import mlc_llm; print(mlc_llm)"Traceback (most recent call last): File "<string>", line 1, in <module> ModuleNotFoundError: No module named 'mlc_llm'
pip list 已安装mlc-llm
更换安装方式
安装环境
-  CMake >= 3.24 
-  Git 
-  Rust and Cargo, required by Hugging Face’s tokenizer 
-  One of the GPU runtimes: -  CUDA >= 11.8 (NVIDIA GPUs) 
-  Metal (Apple GPUs) 
-  Vulkan (NVIDIA, AMD, Intel GPUs) 
 
-  
安装
# clone from GitHub git clone --recursive https://github.com/mlc-ai/mlc-llm.git && cd mlc-llm/ # create build directory mkdir -p build && cd build # generate build configuration python ../cmake/gen_cmake_config.py # build mlc_llm libraries cmake .. && cmake --build . --parallel $(nproc) && cd ..
全局路径
export MLC_LLM_SOURCE_DIR=/path-to-mlc-llm export PYTHONPATH=$MLC_LLM_SOURCE_DIR/python:$PYTHONPATH alias mlc_llm="python -m mlc_llm"