InfoQ

DeepSeek Open-Sources DeepSeek-V3, a 671B Parameter Mixture of Experts LLM

DeepSeek open-sourced DeepSeek-V3, a Mixture-of-Experts (MoE) LLM containing 671B parameters. It was pre-trained on 14.8T tokens using 2.788M GPU hours and outperforms other open-source models on a ...
Fast Company

Curious about DeepSeek but worried about privacy? These apps let you use an LLM without the internet

But thanks to a few innovative and easy-to-use desktop apps, LM Studio and GPT4All, you can bypass both these drawbacks. With the apps, you can run various LLM models on your computer directly. I’ve ...

DeepSeek open sources DSpark, a new framework to speed up LLM inference by up to 85%

DSpark can make decoding faster, but acceptance quality still determines how much speed the system actually realizes.

How DeepSeek’s DSpark Makes LLMs 85% Faster Without Retraining

DeepSeek has introduced DSpark, an open-source speculative decoding method that accelerates LLM inference by up to 85% ...
SiliconANGLE

DeepSeek open-sources new AI model with 671B parameters

Chinese artificial intelligence developer DeepSeek today open-sourced DeepSeek-V3, a new large language model with 671 billion parameters. The LLM can generate text, craft software code and perform ...
Geeky Gadgets

DeepSeek-v2.5 open source LLM performance tested – Beats Claude 3, GPT-4o and Google Gemini

The development of DeepSeek v2.5 involved the fusion of two highly capable models: DeepSeek version 2 0628 and DeepSeek Coder version 2 0724. By combining the strengths of these models, DeepSeek v2.5 ...
InfoQ

DeepSeek Open-Sources DeepSeek-R1 LLM with Performance Comparable to OpenAI's o1 Model

A monthly overview of things you need to know as an architect or aspiring architect. Unlock the full InfoQ experience by logging in! Stay updated with your favorite authors and topics, engage with ...
Dark Reading

LLM Hijackers Quickly Incorporate DeepSeek API Keys

Sophisticated "LLMjacking" operations have obtained stolen access to DeepSeek models, just weeks after their public release. Most recently, researchers from Sysdig observed hyperactive LLMjacking ...
VentureBeat

DeepSeek’s conditional memory fixes silent LLM waste: GPU cycles lost to static lookups

When an enterprise LLM retrieves a product name, technical specification, or standard contract clause, it's using expensive GPU computation designed for complex reasoning — just to access static ...
SDxCentral

DeepSeek looks to offload simple LLM tasks to save billions of parameters

A little over a year after it upended the tech industry, DeepSeek is back with another apparent breakthrough: a means to stop current large language models (LLMs) from wasting computational depth on ...