Kyle Boddy

We are considering the move to Zulip. Is there a guide for this yet? I realize it's four years old, but data migration, while important, is less important than bot...

I am also running into this ingesting a large number of Slack documents/messages. I posted to the danswer slack Support channel as well. 

Logs when using `grep` to check and getting the 5 lines before/after those errors: ``` kyle@danswer:~$ sudo docker logs danswer-stack-background-1 | grep range -C 5 /usr/local/lib/python3.11/site-packages/supervisor/options.py:474: UserWarning: Supervisord is running...

Adding more logs here that also has the duration of the timeout. When I just try to ingest the "general" channel I run into the same problem. ``` [2024-01-04 07:32:57,554:...

This was possibly addressed by this code change: https://github.com/danswer-ai/danswer/pull/910

As far as I can tell using multi-GPU affords zero speedup. The code appears to just round-robin access the GPUs after loading the model into each of them. This is...

 Watch this GIF of `nvtop` being used. The GPUs are just used in round robin fashion, not concurrently. (Only my P40s are visible to the machine)

It could do this via chunking and splitting, then combining; there's logic out there that does it. But I don't believe this library handles it natively. Could make for a...

After running it with a single GPU and comparing, the multiGPU method using DEVICE_INDEX_LIST or similar is definitely used but only reduces memory pressure (by a good amount though, nearly...

I've been using ThreadPoolExecutor() as well to run on 3x Tesla P40 GPUs and still run into `CUDA error: an illegal memory access was encountered` every so often. It almost...