Closed KastanDay closed 2 years ago
@yuanchi2807 @klwuibm - any ideas? Have you seen this before?
HI @KastanDay, thanks for checking out CodeFlare. The error appears to be originated from sklearn and a bug in cython. What version of sklearn are you using? Can you try again with 0.24.+? Thanks.
I'm using the latest stable version of sklearn sklearn.__version__
is 1.0
.
Codeflare version codeflare-0.1.2.dev0
from pip3 install .
Still unsuccessful after a few more attempts:
Same error :/
I wouldn't mind except I get the same error when trying to build my own more sophisticated pipeline as part of my research!
Edit: Of course, the docker container works no problem. Yet, I'd ideally like to contribute to your project, so I need it running locally :)
Thanks, would be happy to work with you to find a solution. Best, Kastan
Hi @KastanDay, thank you for confirming sklearn version to be 1.0. I am also getting the "ValueError: buffer source array is read-only" in my environment after advancing to 1.0. However, after downgrading scikit-learn to 0.24.1, the notebook ran as expected with no exception thrown. It appears to me that sklearn 1.0 broke ray. I am searching ray forum to see if others have reported this issue. For reference, I am running python 3.8.8 and ray 2.0.0.dev0. sklearn 1.0 => read-only exception sklearn 0.24.1 => pass CC: @raghukiran1224 @klwuibm
Hi @KastanDay, turned out this is a recurring bug in pandas. It was reported and investigated by ray developers. See this comment There are still multiple bug reports on its regression in later Pandas releases See
I have tried multiple combinations of pandas and scikit-learn but the only version that works is sklearn 0.24.1. Do you have a strong dependency on sklearn 1.0?
Hi @yuanchi2807, thank you for the detailed investigation with interesting results. I have no dependency to sklearn 1.0, so I will downgrade as you suggest.
Describe the bug I'm trying to run the example notebooks (in
codeflare/notebooks
), and came across this error. The error persisted thru attempts to restart my kernel, entire machine, and re-cloning the repo. Any help, or an explanation of the root cause, is much appreciated!To Reproduce Steps to reproduce the behavior:
notebooks/plot_nca_classification.ipynb
knn_pipeline = rt.select_pipeline(pipeline_fitted, pipeline_fitted.get_xyrefs(node_knn)[0])
RaySystemError: System error: buffer source array is read-only
Full stack trace:
Expected behavior Expected is selecting the pipeline and evaluating its score via a 'SCORE' pipeline.
Desktop
Thank you for any help! I am a University of Illinois at Urbana-Champaign grad student trying to make the most of your work!