Open JorgeVindelAlfageme opened 1 month ago
Here's how the error message looks like:
Which installation method did you use? conda or pip? Also which Windows operating system are you using?
I chose "Install using conda" path. I'm using a 64-bit Windows. I created a second environment using the "pip" method and a CPU-based approach and now I'm able to import Keypoint-MoSeq in Jupyter Notebook, but I would like to know more about how to use my GPU and those systems that depend on a related installation process. Here are my computer details if you want to check them directly. Thank you for your time:
Did you see this note on the installation page and try following it?
Yes, but since it was indicated on my computer that I had the most updated drivers, I thought that that wasn't the problem. Should I get new drivers from some internet platform?
Based on your GPU model and operating system, it looks like a newer driver is available: https://www.nvidia.com/download/driverResults.aspx/224483/en-us/
Thank you for your answer. For now, I'm using the CPU-based version of Keypoint-MoSeq, but I want to check if I can solve this inconvenience with GPU. I'll try to download this new driver you recommended me and I'll write about the results in the next days.
Hi! I'm new using Keypoint-MoSeq. I'm trying to execute it on my computer, but I haven't been able to. I followed the installation steps as they are indicated on: https://keypoint-moseq.readthedocs.io/en/latest/install.html. After using the command: "conda activate keypoint_moseq" and "jupyter notebook", I tried to import Keypoint-MoSeq, and the next error message appeared:
"XlaRuntimeError: UNKNOWN: no kernel image is available for execution on the device in external/org_tensorflow/tensorflow/compiler/xla/stream_executor/cuda/cuda_asm_compiler.cc(64): 'status'".
I guess that, based on what appears on: https://keypoint-moseq.readthedocs.io/en/latest/install.html, it has to do with the GPU version that I'm using. I tried to install CUDA using: "conda install cuda -c nvidia", but the problem wasn't solved. I'm not used to use my GPU, so I don't know how to solve it. Do you guys know what I should do? Thank you.
P. D.: The "nvidia-smi" output from Windows Powershell can be found on the next image: