ComfyUI setup for AMD Radeon GPU on dbox260606 — text-to-video generation
| AGENTS.md | ||
| PROGRESS.md | ||
| README.md | ||
| SETUP.md | ||
| TASKS.md | ||
AMD-comfyui
ComfyUI text-to-video generation on AMD Radeon GPU, running inside dbox260606 (Ubuntu 24.04 distrobox on PC00).
Hardware
| GPU | AMD Radeon RX 7600 (Navi 33, gfx1102, PCI 1002:7480) |
| Host | PC00 — Ubuntu 24.04.3 LTS |
| Container | dbox260606 (Ubuntu 24.04 distrobox) |
| Install path | /mnt/250721_HDD/ComfyUI/ |
| Models path | /mnt/250721_HDD/ComfyUI/models/ |
Quick start (after full setup is complete)
# Enter container
distrobox enter dbox260606
# Activate environment and launch
source /mnt/250721_HDD/ComfyUI/venv/bin/activate
export HSA_OVERRIDE_GFX_VERSION=11.0.0
cd /mnt/250721_HDD/ComfyUI
python3 main.py --listen 0.0.0.0 --port 8188
Access the UI at http://localhost:8188 from any browser on PC00 or the LAN.
Documents
| File | Purpose |
|---|---|
SETUP.md |
Full step-by-step setup plan |
TASKS.md |
Agent task checklist — work through top to bottom |
AGENTS.md |
Agent instructions and RLC rules |
PROGRESS.md |
Session log — append every session |
Model used
LTX Video (ltx-video-2b-v0.9.5) by Lightricks — efficient 2B-parameter video diffusion model.
Text encoder: T5-XXL (fp16).
Source
Based on: AMD AI Tutorials v11.0 — Text-to-video generation with ComfyUI and an AMD Radeon GPU