Unrelated, but wondering if anyone here could recommend a Darktable-ish web-based photo organization app, less focused on editing but supporting tagging, starring, etc.?
...pods could some how mount node{1..5} as a volume, which would have 5 * 200GB ~1TB of space to write to... multiple pods could mount it and read the same data.
Somewhat related: can anyone recommend a simple solution to share each node’s ephemeral disk/“emptyDir” across the cluster? Speed is more important than durability, this is just for a temporary batch job cluster. It’d be ideal if I could stripe across nodes and expose one big volume to all pods (JBOD style)
One thing I'm not sure of is how much of a larger bit of text should go into an embedding? I assume it's a trade off of context and recall, with one word not meaning much semantically, and the whole document being too much to represent with just numbers. Is there a sweet spot (e.g. split by sentence) or am I missing something here?
As far as I know, even with the recent increase, Backblaze B2 is still the most cost efficient option at $6/TB/mo... but I'll be watching replies closely.
I wonder if this could be used to port entire projects/libraries to different languages almost completely automatically? ElasticSearch in Rust? Wondering about the legal/license ramifications of doing so...
Just some anecdata but I'm not seeing such dramatic improvements with YJIT, actually in my limited testing it's much slower...
Ruby 3.2.0, on a M1 MacBook Pro hitting a simple healthcheck route in a production Rails 6.1 app with wrk (`wrk -c8 -t8 -d30s http://localhost:3000/healthcheck`):
- 423 reqs/second with JIT disabled (using 229MB of RAM)
- 74 reqs/second running with `RUBYOPT="--yjit"` (using 385 MB of RAM)
> The Averting the National Threat of Internet Surveillance, Oppressive Censorship and Influence, and Algorithmic Learning by the Chinese Communist Party Act (ANTI-SOCIAL CCP Act)