Qwen Image 2.1 - SD.CPP - terminal scripts

A developer published a set of bash scripts for running Qwen Image 2.1 image editing locally through sd.cpp's sd-cli, pairing the Q4_K_M GGUF diffusion model and its bf16 VAE with a Qwen3-VL-8B-Instruct LLM and vision projector. The scripts cover text-guided edits, multi-region instruction-based edits, and LoRA-driven outpainting at 512x512 and 512x768 with 20 steps and a 6.0 CFG scale, logging elapsed time and opening each result automatically.

| | /bin/bash | | | start time=$ date +%s | | | echo "" | | | CURRENT SEED=$RANDOM | | | OUTPUT FILE=output/i2i image-edit $ date '+%F %H %M %S' $CURRENT SEED.png | | | positive prompt=""" | | | the sloth fur color is now blue | | | """ | | | ./sd-cpp/sd-cli --diffusion-model models/qwen-image-2.1-Q4 K M.gguf \ | | | --vae models/qwen image 2.1 vae bf16.safetensors \ | | | --llm models/Qwen3-VL-8B-Instruct-Q4 0.gguf \ | | | --llm vision models/mmproj-F16 Qwen3-VL-8B-Instruct.gguf \ | | | --cfg-scale 6.0 --steps 20 --sampling-method euler -W 512 -H 512 --diffusion-fa \ | | | --prompt "{$positive prompt}" \ | | | --negative-prompt "" \ | | | -r output/t2i 2026-10-03 14 04 17 11803.png \ | | | --output "$OUTPUT FILE" | | | open the output image in the default associated app. | | | xdg-open "$OUTPUT FILE" | | | end time=$ date +%s | | | elapsed=$ end time - start time | | | echo "" | | | echo "Elapsed time: $elapsed seconds" | | | Wait for the user to press Enter | | | echo "" | | | echo "+-----------------------------------------------+" | | | echo "Press ENTER to Exit." | | | read -r | | | /bin/bash | | | start time=$ date +%s | | | echo "" | | | CURRENT SEED=$RANDOM | | | OUTPUT FILE=output/i2i $ date '+%F %H %M %S' $CURRENT SEED.png | | | positive prompt=""" | | | Remove the metal watch in the blue circle, change the hair in the red circle to black, and replace the area in the green circle with gray short-sleeved linen pajamas. | | | """ | | | ./sd-cpp/sd-cli --diffusion-model models/qwen-image-2.1-Q4 K M.gguf \ | | | --vae models/qwen image 2.1 vae bf16.safetensors \ | | | --llm models/Qwen3-VL-8B-Instruct-Q4 0.gguf \ | | | --llm vision models/mmproj-F16 Qwen3-VL-8B-Instruct.gguf \ | | | --cfg-scale 6.0 --steps 20 --sampling-method euler -W 512 -H 512 --diffusion-fa \ | | | --prompt "{$positive prompt}" \ | | | --negative-prompt "" \ | | | -r output/example-18.png \ | | | --output "$OUTPUT FILE" | | | open the output image in the default associated app. | | | xdg-open "$OUTPUT FILE" | | | end time=$ date +%s | | | elapsed=$ end time - start time | | | echo "" | | | echo "Elapsed time: $elapsed seconds" | | | Wait for the user to press Enter | | | echo "" | | | echo "+-----------------------------------------------+" | | | echo "Press ENTER to Exit." | | | read -r | | | /bin/bash | | | start time=$ date +%s | | | echo "" | | | CURRENT SEED=$RANDOM | | | OUTPUT FILE=output/i2i outpaint $ date '+%F %H %M %S' $CURRENT SEED.png | | | positive prompt=""" | | | <lora:qwen-image-2.1-outpaint-v2:1.0 | | | Outpaint the image: replace the solid gray areas with a seamless continuation of the scene, keeping the existing picture unchanged. | | | Scene: <description | | | """ | | | ./sd-cpp/sd-cli --diffusion-model models/qwen-image-2.1-Q4 K M.gguf \ | | | --vae models/qwen image 2.1 vae bf16.safetensors \ | | | --llm models/Qwen3-VL-8B-Instruct-Q4 0.gguf \ | | | --llm vision models/mmproj-F16 Qwen3-VL-8B-Instruct.gguf \ | | | --lora-model-dir ./models/lora/ \ | | | --cfg-scale 6.0 --steps 20 --sampling-method euler -W 512 -H 768 --diffusion-fa \ | | | --prompt "{$positive prompt}" \ | | | --negative-prompt "" \ | | | -r output/bladerunner ana-de-armas joi hologram-bridge 01.jpg \ | | | --output "$OUTPUT FILE" | | | open the output image in the default associated app. | | | xdg-open "$OUTPUT FILE" | | | end time=$ date +%s | | | elapsed=$ end time - start time | | | echo "" | | | echo "Elapsed time: $elapsed seconds" | | | Wait for the user to press Enter | | | echo "" | | | echo "+-----------------------------------------------+" | | | echo "Press ENTER to Exit." | | | read -r | | | /bin/bash | | | start time=$ date +%s | | | echo "" | | | CURRENT SEED=$RANDOM | | | base image=output/example-18.png | | | style image=output/i2i 2026-10-04 20 15 08 28547.png | | | OUTPUT FILE=output/i2i $ date '+%F %H %M %S' $CURRENT SEED.png | | | positive prompt=""" | | | apply <image2 art style to <image1 | | | """ | | | ./sd-cpp/sd-cli --diffusion-model models/qwen-image-2.1-Q4 K M.gguf \ | | | --vae models/qwen image 2.1 vae bf16.safetensors \ | | | --llm models/Qwen3-VL-8B-Instruct-Q4 0.gguf \ | | | --llm vision models/mmproj-F16 Qwen3-VL-8B-Instruct.gguf \ | | | --cfg-scale 6.0 --steps 20 --sampling-method euler --diffusion-fa \ | | | --prompt "{$positive prompt}" \ | | | --negative-prompt "" \ | | | -r $base image \ | | | -r $style image \ | | | --output "$OUTPUT FILE" | | | open the output image in the default associated app. | | | xdg-open "$OUTPUT FILE" | | | end time=$ date +%s | | | elapsed=$ end time - start time | | | echo "" | | | echo "Elapsed time: $elapsed seconds" | | | Wait for the user to press Enter | | | echo "" | | | echo "+-----------------------------------------------+" | | | echo "Press ENTER to Exit." | | | read -r | | | /bin/bash | | | start time=$ date +%s | | | echo "" | | | CURRENT SEED=$RANDOM | | | OUTPUT FILE=output/i2i $ date '+%F %H %M %S' $CURRENT SEED.png | | | positive prompt=""" | | | This is an RGBA image with transparency. make the white background transparent. The image has alpha channel and the background is transparent. | | | """ | | | ./sd-cpp/sd-cli --diffusion-model models/qwen-image-2.1-Q4 K M.gguf \ | | | --vae models/qwen image 2.1 vae bf16.safetensors \ | | | --llm models/Qwen3-VL-8B-Instruct-Q4 0.gguf \ | | | --llm vision models/mmproj-F16 Qwen3-VL-8B-Instruct.gguf \ | | | --cfg-scale 6.0 --steps 20 --sampling-method euler -W 512 -H 512 --diffusion-fa \ | | | --prompt "{$positive prompt}" \ | | | --negative-prompt "" \ | | | -r output/i2i 2026-10-04 20 15 08 28547.png \ | | | --output "$OUTPUT FILE" | | | open the output image in the default associated app. | | | xdg-open "$OUTPUT FILE" | | | end time=$ date +%s | | | elapsed=$ end time - start time | | | echo "" | | | echo "Elapsed time: $elapsed seconds" | | | Wait for the user to press Enter | | | echo "" | | | echo "+-----------------------------------------------+" | | | echo "Press ENTER to Exit." | | | read -r | | | /bin/bash | | | start time=$ date +%s | | | echo "" | | | CURRENT SEED=$RANDOM | | | OUTPUT FILE=output/i2i $ date '+%F %H %M %S' $CURRENT SEED.png | | | input prompt=""" | | | a cartoon sloth mascot waving, flat vector illustration, bright colours | | | """ | | | do not change positive prompt as the model was trained to accept this schema. | | | positive prompt=""" | | | This is an RGBA image with transparency. {$input prompt}. The image has alpha channel and the background is transparent. | | | """ | | | ./sd-cpp/sd-cli --diffusion-model models/qwen-image-2.1-Q4 K M.gguf \ | | | --vae models/qwen image 2.1 vae bf16.safetensors \ | | | --llm models/Qwen3-VL-8B-Instruct-Q4 0.gguf \ | | | --llm vision models/mmproj-F16 Qwen3-VL-8B-Instruct.gguf \ | | | --cfg-scale 6.0 --steps 20 --sampling-method euler -W 512 -H 512 --diffusion-fa \ | | | --prompt "{$positive prompt}" \ | | | --negative-prompt "" \ | | | --output "$OUTPUT FILE" | | | open the output image in the default associated app. | | | xdg-open "$OUTPUT FILE" | | | end time=$ date +%s | | | elapsed=$ end time - start time | | | echo "" | | | echo "Elapsed time: $elapsed seconds" | | | Wait for the user to press Enter | | | echo "" | | | echo "+-----------------------------------------------+" | | | echo "Press ENTER to Exit." | | | read -r | | | /bin/bash | | | start time=$ date +%s | | | echo "" | | | CURRENT SEED=$RANDOM | | | OUTPUT FILE=output/i2i $ date '+%F %H %M %S' $CURRENT SEED.png | | | positive prompt=""" | | | a cartoon sloth mascot waving, flat vector illustration, bright colours. | | | """ | | | ./sd-cpp/sd-cli --diffusion-model models/qwen-image-2.1-Q4 K M.gguf \ | | | --vae models/qwen image 2.1 vae bf16.safetensors \ | | | --llm models/Qwen3-VL-8B-Instruct-Q4 0.gguf \ | | | --llm vision models/mmproj-F16 Qwen3-VL-8B-Instruct.gguf \ | | | --cfg-scale 6.0 --steps 20 --sampling-method euler -W 1024 -H 1024 --diffusion-fa \ | | | --prompt "{$positive prompt}" \ | | | --negative-prompt "" \ | | | --output "$OUTPUT FILE" | | | open the output image in the default associated app. | | | xdg-open "$OUTPUT FILE" | | | end time=$ date +%s | | | elapsed=$ end time - start time | | | echo "" | | | echo "Elapsed time: $elapsed seconds" | | | Wait for the user to press Enter | | | echo "" | | | echo "+-----------------------------------------------+" | | | echo "Press ENTER to Exit." | | | read -r |