[GH-ISSUE #2139] Use the Kontext model in AMD gpu 9070xt, always use shared VRAM #875

Closed
opened 2026-10-27 21:23:24 +00:00 by bsakaguchi · 0 comments
Owner

Originally created by @fangwen1102 on GitHub (Sep 7, 2025).
Original GitHub issue: https://github.com/Comfy-Org/ComfyUI-Manager/issues/2139

No matter it is FP8 or GGUF Q4 model, it always uses 14G VRAM and 2.5G shared VRAM.
comfyui Official flux Kontext dev example workflow, only the model or gguf is modified.
python 3.13
rocm 7.0.0rc20250906
rocm-sdk-core 7.0.0rc20250906
rocm-sdk-devel 7.0.0rc20250906
rocm-sdk-libraries-gfx120X-all 7.0.0rc20250906
torch 2.9.0a0+rocm7.0.0rc20250906
torchaudio 2.8.0a0+rocm7.0.0rc20250906
torchvision 0.24.0a0+rocm7.0.0rc20250906

Image
Originally created by @fangwen1102 on GitHub (Sep 7, 2025). Original GitHub issue: https://github.com/Comfy-Org/ComfyUI-Manager/issues/2139 No matter it is FP8 or GGUF Q4 model, it always uses 14G VRAM and 2.5G shared VRAM. comfyui Official flux Kontext dev example workflow, only the model or gguf is modified. python 3.13 rocm 7.0.0rc20250906 rocm-sdk-core 7.0.0rc20250906 rocm-sdk-devel 7.0.0rc20250906 rocm-sdk-libraries-gfx120X-all 7.0.0rc20250906 torch 2.9.0a0+rocm7.0.0rc20250906 torchaudio 2.8.0a0+rocm7.0.0rc20250906 torchvision 0.24.0a0+rocm7.0.0rc20250906 <img width="2256" height="1100" alt="Image" src="https://github.com/user-attachments/assets/65ad632f-a526-4b7a-a15b-4fa749a8ac12" />
Sign in to join this conversation.
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
Comfy-Org/ComfyUI-Manager#875
No description provided.