Mismatching CLIP model?

[unrealfuzzy]

Hi, ran into an issue with running 5_MAGnet_Generate_Modern_Paintings.ipynb
Exact problem is this:

import clip
from PIL import Image
device = "cpu"
model, preprocess = clip.load("ViT-B/32", device=device)
...
input_resolution = model.input_resolution.item()

which produces this:

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/usr/local/lib64/python3.7/site-packages/torch/nn/modules/module.py", line 1131, in __getattr__
    type(self).__name__, name))
AttributeError: 'CLIP' object has no attribute 'input_resolution'

So seem like current CLIP implementation that was pulled from GitHub doesn't have input_resolution method, and it's not entirely clear what substitute to use. I tried a (512, 512) touple along with just 512 integer and it doesn't seem to work.

Could you please help me out on this one?

[robgon-art]

Hello, thanks for flagging this issue. I will take a look at it. In the meantime, you can try setting input_resolution to a simple integer, 224.