pre-trained stable diffusion model taking 4min to generate an image

[HEGADE]

hi , I'm trying to generate an image from a prompt using the Stable-Diffusion-v1-5 model,and I'm using NVidia GeForce GTX 1650 , to create an image it's taking 4 min, is it normal? or I did something wrong?, how can I improve the performance

[haofanwang]

(1) make sure you do put your model on GPU correctly (2) use float16 (3) set smaller output resolution or decrease steps

[HEGADE]

Okay tq

[sayakpaul]

@haofanwang thanks for sharing your thoughts.

@HEGADE could you close the issue if https://github.com/huggingface/diffusers/issues/2411#issuecomment-1435512428 resolved your problem?

[HEGADE]

issues resolved thank you for the support