search

comfyanonymous / ComfyUI

The most powerful and modular diffusion model GUI, api and backend with a graph/nodes interface.
https://www.comfy.org/
GNU General Public License v3.0
50.75k stars 5.33k forks source link

Unable to generate images #39

Closed vararth closed 2 months ago

vararth commented 1 year ago

Getting this in powershell -

ValueError: The device_map provided does not give any device for the following parameters: betas, alphas_cumprod, alphas_cumprod_prev, sqrt_alphas_cumprod, sqrt_one_minus_alphas_cumprod, log_one_minus_alphas_cumprod, sqrt_recip_alphas_cumprod, sqrt_recipm1_alphas_cumprod, posterior_variance, posterior_log_variance_clipped, posterior_mean_coef1, posterior_mean_coef2

comfyanonymous commented 1 year ago

Can you give me more information about your system like which GPU you have and how you are running comfyUI?

vararth commented 1 year ago

I have a Nvidia 3050, 4GB running on a Lenovo Ideapad. I have automatic1111 already installed so I went to the comfyui folder, started up powershell, ran the Activate.ps1 script and in the models folder, I have waifu diffusion ckpt loaded. Once the url showed up, the UI loaded up and I did not make any changes just to do a dry run of the tool, but nothing was generated after I pressed the button.

comfyanonymous commented 1 year ago

It looks like the a1111 ui venv has an outdated version of accelerate. I use that package for the lowvram mode in my UI which gets automatically enabled on your GPU.

Try: pip install accelerate --upgrade