search

holoviz / lumen

Illuminate your data.
https://lumen.holoviz.org
BSD 3-Clause "New" or "Revised" License
177 stars 20 forks source link

Tweak local model settings #780

Open ahuang11 opened 5 days ago

ahuang11 commented 5 days ago

Pins version and makes minor tweaks to config.

Seems promising with Qwen; couldn't get it running with Mistral 7B

codecov[bot] commented 5 days ago

Codecov Report

Attention: Patch coverage is 0% with 25 lines in your changes missing coverage. Please review.

Project coverage is 60.94%. Comparing base (2c83a1b) to head (1c86fd3).

Files with missing lines Patch % Lines
lumen/ai/coordinator.py 0.00% 23 Missing :warning:
lumen/ai/llm.py 0.00% 2 Missing :warning:
Additional details and impacted files ```diff @@ Coverage Diff @@ ## main #780 +/- ## ========================================== - Coverage 60.98% 60.94% -0.04% ========================================== Files 103 103 Lines 12544 12552 +8 ========================================== Hits 7650 7650 - Misses 4894 4902 +8 ```

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

ahuang11 commented 9 hours ago

Prior to this, since Llama.invoke isn't truly async, there was no loading indicator and it was confusing whether Lumen was doing anything, but this PR changes that:

image image