search

basujindal / stable-diffusion

Optimized Stable Diffusion modified to run on lower GPU VRAM
Other
3.14k stars 469 forks source link

img2img output not available when running with docker #163

Open mbwgh opened 1 year ago

mbwgh commented 1 year ago

If you run e.g.

docker compose run sd python optimizedSD/optimized_txt2img.py --precision full --prompt "A monkey programming code on a computer. The monkey looks very smart." --H 512 --W 512 --seed 27 --n_iter 2 --n_samples 5 --ddim_steps 50

the output becomes available in the volume specified in the docker-compose.yml.

Not so when trying the same when you use optimized_img2img.py instead.

I believe it is due to

RUN ln -s /data /root/stable-diffusion/models/ldm/stable-diffusion-v1 \
│&& mkdir -p /output /root/stable-diffusion/outputs \
│&& ln -s /output /root/stable-diffusion/outputs/txt2img-samples

in the Dockerfile.