search

rapidsai / jupyterlab-nvdashboard

A JupyterLab extension for displaying dashboards of GPU usage.
BSD 3-Clause "New" or "Revised" License
579 stars 77 forks source link

Support bokeh 3 #141

Open jacobtomlinson opened 1 year ago

jacobtomlinson commented 1 year ago

Looks like things are a bit broken now with bokeh 3. See #139 #131 for info.

We should update things here to work nicely with Bokeh 3.

cc @bryevdv

samarjeet commented 1 year ago

Hi, any update on this? I get the error : pynvml.nvml.NVMLError_DriverNotLoaded: Driver Not Loaded

bryevdv commented 1 year ago

@samarjeet I don't see how that could be related to bokeh in any way

samarjeet commented 1 year ago

@bryevdv Ah, my mistake. Definitely not the right issue. Feel free to delete it from this thread and I can move the question to another relevant one. My error is similar to #139 and that thread was closed with a note that this thread #141 is going to handle the issue.

Anyways, the problem is that server is trying to load jupyterlab-nvdashboard on a non-gpu machine. Only if I comment out from . import gpu here, am I able to launch jupyterlab.

paravatha commented 1 year ago

Hello, I am seeing issue https://github.com/rapidsai/jupyterlab-nvdashboard/issues/139 when I use bokeh==3.0.2 It works when I use bokeh=2.4.3

bryevdv commented 1 year ago

There's like 4-5 different descriptions of problems spread across three issues at this point. I don't really use Jupyter much (I actively avoid it to the extent possible). Can someone provide, in one place, a complete concise set of steps to reproduce a bokeh-related problem (noting your audience: someone not familiar with this nvdashboard tool, or with using or running it—start from the beginning), and a concise description of that problem?