To add for me (and claude assisted but tuned myself), here is for 3060 12gb + 5950x 128 gb ram without image support: https://pastebin.com/HsAUBGSY https://past

To add for me (and claude assisted but tuned myself), here is for 3060 12gb + 5950x 128 gb ram without image support: https://pastebin.com/HsAUBGSY https://pastebin.com/HsAUBGSY Here is the chat template I used (and renamed to qwen.jinja): https://huggingface.co/froggeric/Qwen-Fixed-Chat-Templates/raw/main/chat_template.jinja https://huggingface.co/froggeric/Qwen-Fixed-Chat-Templates/r... For image support, drop to NGL 24 and add: --mmproj-url https://huggingface.co/unsloth/Qwen3.8-27B-GGUF/resolve/main/mmproj-F16.gguf \ --no-mmproj-offload \ --image-min-tokens 1024 \ And drop the --no-mmproj