search

jina-ai / discoart

🪩 Create Disco Diffusion artworks in one line
Other
3.84k stars 249 forks source link

Runtime Error with secondary model off #120

Closed dillfrescott closed 2 years ago

dillfrescott commented 2 years ago

RuntimeError Traceback (most recent call last) in () 5 n_batches=1, 6 use_secondary_model=False, ----> 7 steps=250, 8 )

19 frames /usr/local/lib/python3.7/dist-packages/torch/nn/modules/conv.py in _conv_forward(self, input, weight, bias) 452 _pair(0), self.dilation, self.groups) 453 return F.conv2d(input, weight, bias, self.stride, --> 454 self.padding, self.dilation, self.groups) 455 456 def forward(self, input: Tensor) -> Tensor:

RuntimeError: Unable to find a valid cuDNN algorithm to run convolution

hanxiao commented 2 years ago
  1. what is your VRAM?
  2. what is your DiscoArt version, try update to latest pip install -U discoart
  3. use_secondary_model=False use more memory than use_secondary_model=True (that's why it is default)

this is the result on default create() on 0.10.4 , so less than 12GB

image

and this is when secondary model is off, <20GB

from discoart import create

create(use_secondary_model=False)

image

dillfrescott commented 2 years ago

Vram is 16 GB. Maybe thats why. Thank you for the info!