Demetris Roumis
Demetris Roumis
@ahuang11 I updated the argparser so that running `lumen-ai --help` or `lumen-ai serve --help` actually returned the relevant information, rather than the panel/bokeh commands.
Since there is a way to upload data from the GUI, should we also make it possible to just run `lumen-ai` without any other args? I think this might involve...
I'm guessing you are referring to serving a separate app script in a file; I'm not sure how that applies to the CLI context. But no worries. Could be something...
I'll give it a quick shot
Was quite a bit more than a 'quick' shot.. but I got it working.. maybe not the prettiest solution. https://github.com/user-attachments/assets/0c476439-d577-4f5e-bd35-3d3bd02e1b1c
I'm having to use the following on M1 `CMAKE_ARGS="-DLLAMA_METAL=on -DCMAKE_OSX_ARCHITECTURES=arm64" FORCE_CMAKE="1" pip install 'llama-cpp-python'`
I think I might be confused, but assuming the 'after' below is the proposed API, so far it saves a bit of time by not having to do a groupby,...
Got it. It would be nice if there was a convenient way to report timing of the initial and subsequent paint. Hopefully soon! While a dict of df's is so...
This approach may not be feasible for dozens of lines. At some point the resulting imagestack (1 image per line) grows beyond what is reasonably performant
Hi @mattpap , what technical fiddling is needed for this?