flash attention (-fa 1) and no-mmap - model.ini #3

Open
opened 2026-08-15 13:08:39 +01:00 by apb · 0 comments
Owner

<<<
⚠️ IMPORTANT: Always use flash attention (-fa 1) and no-mmap (--no-mmap) on Strix Halo to avoid crashes/slowdowns.

Server Mode (API):

llama-server -m models/qwen3-coder-30B-A3B/BF16/Qwen3-Coder-30B-A3B-Instruct-BF16-00001-of-00002.gguf
-c 8192 -ngl 999 -fa 1 --no-mmap
<<<https://github.com/kyuz0/amd-strix-halo-toolboxes#4-run-inference

https://github.com/kyuz0/amd-strix-halo-toolboxes/blob/main/docs/models.ini.example

<<< ⚠️ IMPORTANT: Always use flash attention (-fa 1) and no-mmap (--no-mmap) on Strix Halo to avoid crashes/slowdowns. Server Mode (API): llama-server -m models/qwen3-coder-30B-A3B/BF16/Qwen3-Coder-30B-A3B-Instruct-BF16-00001-of-00002.gguf \ -c 8192 -ngl 999 -fa 1 --no-mmap <<<https://github.com/kyuz0/amd-strix-halo-toolboxes#4-run-inference https://github.com/kyuz0/amd-strix-halo-toolboxes/blob/main/docs/models.ini.example
Sign in to join this conversation.
No labels
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
apb/ai-code#3
No description provided.