envolution

@cass3451 I don't think this should be closed since there is no current solution for numpy 2

understood, thank you!

I tried without (shown below) -std=c++11 and with 14,17,20,23 alternatively without luck passing the "error: user-defined literal operator not found" ``` $ make cuda_v12 CUDA_12_PATH=/opt/cuda /usr/bin/ccache /opt/cuda/bin/nvcc -c -Xcompiler -fPIC...

I can confirm that my posts are always within a few minutes of a fresh clone: ``` sh-5.2$ date Wed 11 Dec 2024 11:49:51 AM EST sh-5.2$ git clone https://github.com/ollama/ollama.git...

it's possible llama does not support my architecture as it's pretty dated. I came across an old report I had on llama.cpp and noticed you had commented as well. https://github.com/ggerganov/llama.cpp/issues/4395...

here's a llama.cpp compilation log showing the same error - this issue isn't rooted in ollama ``` $ CUDAToolkit_ROOT=/opt/cuda CUDACXX=/opt/cuda/bin/nvcc cmake -B build -DGGML_CUDA=ON -- ccache found, compilation results will...

it was yes - i was able to compile by updating with this commit https://aur.archlinux.org/cgit/aur.git/commit/?h=ollama-cuda-git&id=1bae2d23dba06da3c4b0a8e7e7f2f1a37c723ca3

https://github.com/Keeper-Security/Commander/releases/tag/v17.0.0 tagged pre-

I see the release has now been deleted, breaking package management downstream again. Is it not safe to assume pre-releases will not be replaced/deleted after being published? I can understand...