DEV Community

maruhana profile picture

maruhana

Building small things with smart glasses, M5Stack and wearables. I mainly write in Japanese on Zenn.

Joined Joined on 
I Cut Dedicated GPU Memory to 4GB, and a Local LLM Still Ran at 90% Speed
maruhana profile

I Cut Dedicated GPU Memory to 4GB, and a Local LLM Still Ran at 90% Speed

#ai #llm #llamacpp #nvidia
Comments
15 min read

Want to connect with maruhana?

Create an account to connect with maruhana. You can also sign in below to proceed if you already have an account.

Create Account
Already have an account? Sign in
A 120B Model Ran About 4x Faster Than a 27B One: 8 Local LLMs on a Surface Laptop Ultra
maruhana profile

A 120B Model Ran About 4x Faster Than a 27B One: 8 Local LLMs on a Surface Laptop Ultra

#ai #llm #llamacpp #localllm
Comments 4
13 min read
Moving ComfyUI from Intel Arc to Arm WSL2: The Same requirements.txt Didn't Install the Same Versions
maruhana profile

Moving ComfyUI from Intel Arc to Arm WSL2: The Same requirements.txt Didn't Install the Same Versions

#ai #comfyui #pytorch #machinelearning
Comments 2
16 min read
Ollama vs. llama.cpp on the Same GGUF: The 15% Speed Gap Was the Defaults, Not the Engine
maruhana profile

Ollama vs. llama.cpp on the Same GGUF: The 15% Speed Gap Was the Defaults, Not the Engine

#ai #llm #ollama #llamacpp
Comments 5
7 min read
Rokid Glasses Can Record 4 Microphone Channels, But Only If You Ask for Exactly 4
maruhana profile

Rokid Glasses Can Record 4 Microphone Channels, But Only If You Ask for Exactly 4

#android #java #audio #smartglasses
Comments
13 min read
loading...