Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!


Shells Virtual Desktop
BMail.ag - Secure Email Service
Server.net
CPLicense.net
VPS Server
Buy VPN
Vultr
VMs for AI
HostDare
HostDare
ReliableSite White-Label Dedicated Hosting for Resellers
InterServer VPS
BMail.ag - Secure Email Service
Best VPN
High-Performance Bare Metal Server Solutions
Karvl.com
Server Mania Cloud Hosting
DataWagon Hosting
AlphaVPS Hosting
Evoxt.com
Clouvider
VPS Hosting with NVMe
Residential IPs in the US & 4G Mobile Proxies in EU & US with Unlimited Bandwidth
ReliableSite White-Label Dedicated Hosting for Resellers
Rabisu - Hosting Solutions
Shells Virtual Desktop
New on LowEndTalk? Please Register and read our Community Rules.

All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.

AI server GPU recommendations

2»

Comments

  • @pompompurin said:

    @professorparabellum said:
    i want to build a machine to run a local LLM. i run a minecraft server that sees around 1200 players peak and we get somewhere in the ballpark of 1 million messages daily which is impossible to moderate. We have an AI chat filter but it runs on gemini, and even though they have the cheapest tokens we would still eat through an astronomical amount on a daily basis.
    What GPU should i go with for this? I was looking at the Nvidia T4 because it has very low TDP for how powerful it is and its around $1k. They also have the L4 which is twice as fast and is rated for the same TDP as the T4 but its alot more expensive.

    i'd say switch to deepseek

    No.

    @SwimmingFrog said:
    At stated above, you don't need a LLM for this. Yours is a binary classification task (predict toxic or not toxic). You don't need to generate any text, which is the expensive part of such models.

    Check out for toxic classification models at HuggingFace https://huggingface.co/models?sort=downloads&search=toxic

    With models based on BERT and optimized, you can classify a chat message in ≈1ms with a T4 or 10ms in a CPU.

    This is the only right answer.

Sign In or Register to comment.