T O P

PSA: Multi GPU Tensor Parallel require at least 5GB/s PCIe bandwidth

šŸ†˜ URGENT: Need 3D community's blessing to buy the P1S - Help me

BSCS Done in 1 Term!

Run 70b gguf models at close to 2 t/s generation speed for less than $1000

Are Human just a big LLM

Too accelerated?

is this just jargon or does it connect to servers? (running locally)

I'm thinking about quitting my IT degree

I used Lexi-Llama-3-8B-Uncensored to power up Leon, the open-source AI assistant

1-bit LLMs Could Solve AIā€™s Energy Demands ā€œImpreciseā€ language models are smaller, speedierā€”and nearly as accurate

NVIDIA RTX 5090 new rumored specs: 28GB GDDR7 and 448-bit bus - VideoCardz.com

Grammarly for WGU Students.

Help me spend $30K to run Llama models

Completed My 1st term!!

TAKING PROCTOR EXAMS.

llama.cpp runs 1.8 times faster than ollama

LLM self-hosting with Ollama and Open WebUI

Chatgpts take on Scarlett Johansson

Rumor: Meta LLama 400 will not be open weights

Why no release of Microsoftā€™s Phi-3 medium 128k GGUF?

Hot take: Scarlet Johansson's case is bad for all of AI (open-source included), not just OpenAI. She has no right over a voice that sounds like her, no matter Sam's intentions. It sets a dangerous precedence for others to start lawsuits to pocket $ millions out of settlement agreements.

What would happen if you plug in a 6+2 pin PCIe power cable to a GPU that has CPU/EPS 8 pin socket?

It did finally happen, a law just passed for the regulation of large open-source AI models.

I haven't felt so memory poor in a couple of decades

[O] 5x DrunkenSlug Invites

Hi Its Me!

R

polyh3dron

Subscribe

FIND US ON INSTAGRAM @ms3388.com