Skip to content
View unixyo's full-sized avatar

Block or report unixyo

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
unixyo/README.md

Hi, I'm unixyo

I run large local AI models on one AMD Radeon RX 6900 XT (RDNA2, gfx1030) and write down what actually works, with measurements and limits.

RX 6900 XT 한 장으로 로컬 AI를 돌리고, 실제로 된 설정과 측정값을 정리해 둡니다.

RDNA2 / gfx1030 recipes

Repository What it is
strata-rdna2-gfx1030-recipe Qwen3.8-Flash-Next on Strata with one RX 6900 XT + 64 GB RAM: ~1,200 tok/s prefill, 50-64 tok/s decode, 256K context. Settings, patches, measurements and what did not help.
rdna2-gfx1030-imagegen-recipes ComfyUI setups for Qwen-Image 2.1, MOSS-Image and MiniMax H3 on the same card.

Tools

Repository What it is
openai-daily-free-gate A local proxy that stops OpenAI API requests before they exceed the daily free-token limits of the data-sharing program.

Upstream reports

Ubuntu · ROCm · ComfyUI · llama.cpp · Strata

Pinned Loading

  1. openai-daily-free-gate openai-daily-free-gate Public

    Unofficial local proxy that stops OpenAI API requests before they exceed the daily free-token limits of the data-sharing program

    Python

  2. rdna2-gfx1030-imagegen-recipes rdna2-gfx1030-imagegen-recipes Public

    ComfyUI recipes for Qwen-Image 2.1, MOSS-Image and MiniMax H3 on one RX 6900 XT (RDNA2, gfx1030) + 64 GB RAM

    Python

  3. strata-rdna2-gfx1030-recipe strata-rdna2-gfx1030-recipe Public

    Strata (Qwen3.8-Flash-Next) on one RX 6900 XT (RDNA2, gfx1030) + 64 GB RAM: settings, patches, measurements