The model weights were updated 17 hours ago and now loading them causes the error

[KyriaAnnwyn]

Error message when loading new weights from lambdalabs/sd-image-variations-diffusers:

AttributeError: module transformers has no attribute CLIPImageProcessor

[KyriaAnnwyn]

updated transformers to 4.25.0.dev0

With this version I get the following error:
TypeError: getattr(): attribute name must be string when loading

[generalsvr]

Same error:

AttributeError: module transformers has no attribute CLIPImageProcessor

Any solution?

[KyriaAnnwyn]

@artyemk this one with CLIPImageProcessor is solved by installing 4.25.0.dev0 transformers version from transformers repo. In the nets configs they mention that they work with this version, but the next error with getattr() i haven't solved yet

[KyriaAnnwyn]

Other solution is to checkout previous version of nets configs and weights

[generalsvr]

@KyriaAnnwyn I tried to solve this issue by running

pip install --upgrade git+https://github.com/huggingface/transformers.git

But now I get another error:

AttributeError: 'CLIPVisionModelWithProjection' object has no attribute 'get_image_features'

[KyriaAnnwyn]

@artyemk seems like the version of transformers you got is 4.24. Check this please

[mkabatek]

I'm getting the same error. Trying to install 4.25.0.dev0 but i'm not sure how. Any advice?

(lambda-diffusers) PS C:\Users\Landon\stable_diffusion\lambda-diffusers> pip install -Iv transformers==4.25.0.dev0
Using pip 22.3.1 from C:\Users\Landon\anaconda3\envs\lambda-diffusers\lib\site-packages\pip (python 3.7)
ERROR: Could not find a version that satisfies the requirement transformers==4.25.0.dev0 (from versions: 0.1, 2.0.0, 2.1.0, 2.1.1, 2.2.0, 2.2.1, 2.2.2, 2.3.0, 2.4.0, 2.4.1, 2.5.0, 2.5.1, 2.6.0, 2.7.0, 2.8.0, 2.9.0, 2.9.1, 2.10.0, 2.11.0, 3.0.0, 3.0.1, 3.0.2, 3.1.0, 3.2.0, 3.3.0, 3.3.1, 3.4.0, 3.5.0, 3.5.1, 4.0.0rc1, 4.0.0, 4.0.1, 4.1.0, 4.1.1, 4.2.0, 4.2.1, 4.2.2, 4.3.0rc1, 4.3.0, 4.3.1, 4.3.2, 4.3.3, 4.4.0, 4.4.1, 4.4.2, 4.5.0, 4.5.1, 4.6.0, 4.6.1, 4.7.0, 4.8.0, 4.8.1, 4.8.2, 4.9.0, 4.9.1, 4.9.2, 4.10.0, 4.10.1, 4.10.2, 4.10.3, 4.11.0, 4.11.1, 4.11.2, 4.11.3, 4.12.0, 4.12.1, 4.12.2, 4.12.3, 4.12.4, 4.12.5, 4.13.0, 4.14.0, 4.14.1, 4.15.0, 4.16.0, 4.16.1, 4.16.2, 4.17.0, 4.18.0, 4.19.0, 4.19.1, 4.19.2, 4.19.3, 4.19.4, 4.20.0, 4.20.1, 4.21.0, 4.21.1, 4.21.2, 4.21.3, 4.22.0, 4.22.1, 4.22.2, 4.23.0, 4.23.1, 4.24.0)
ERROR: No matching distribution found for transformers==4.25.0.dev0

[mkabatek]

@artyemk this one with CLIPImageProcessor is solved by installing 4.25.0.dev0 transformers version from transformers repo. In the nets configs they mention that they work with this version, but the next error with getattr() i haven't solved yet

I was able to solve installing transformers with this command

pip install --upgrade "git+https://github.com/huggingface/[email protected]"

However now I am also getting the error.

Traceback (most recent call last):
  File ".\im2im.py", line 11, in <module>
    pipe = StableDiffusionImageEmbedPipeline.from_pretrained("lambdalabs/sd-image-variations-diffusers")
  File "C:\Users\Landon\anaconda3\envs\lambda-diffusers\lib\site-packages\diffusers\pipeline_utils.py", line 373, in from_pretrained
    load_method = getattr(class_obj, load_method_name)
TypeError: getattr(): attribute name must be string

[hniksoleimani]

install the following version of transformers:

pip install transformers==4.19.2

run the following script:

from pathlib import Path
from lambda_diffusers import StableDiffusionImageEmbedPipeline
from PIL import Image
import torch
device = "cuda" if torch.cuda.is_available() else "cpu"
pipe = StableDiffusionImageEmbedPipeline.from_pretrained(
    "lambdalabs/sd-image-variations-diffusers",
    revision="273115e88df42350019ef4d628265b8c29ef4af5",
    )
pipe = pipe.to(device)
im = Image.open("your/input/image/here.jpg")
num_samples = 4
image = pipe(num_samples*[im], guidance_scale=3.0)
image = image["sample"]
base_path = Path("outputs/im2im")
base_path.mkdir(exist_ok=True, parents=True)
for idx, im in enumerate(image):
    im.save(base_path/f"{idx:06}.jpg")

[generalsvr]

I solved this issue by initializing pipe like this:

pipe = StableDiffusionImageEmbedPipeline.from_pretrained(
 "lambdalabs/sd-image-variations-diffusers",
 revision="273115e88df42350019ef4d628265b8c29ef4af5",
 )

[KyriaAnnwyn]

@artyemk @hniksoleimani this makes you use old version of weights, but more interesting to test the new ones!

[mkabatek]

@hniksoleimani @artyemk

I solved this issue by initializing pipe like this:

pipe = StableDiffusionImageEmbedPipeline.from_pretrained(
 "lambdalabs/sd-image-variations-diffusers",
 revision="273115e88df42350019ef4d628265b8c29ef4af5",
 )

Thank you this works, however now i'm curious is there a way to run this with a lower memory GPU? for example I see people use .half(), or fp16 how would I do that using this script?

Thanks again for your feedback. I am running on Nvidia 3070 with 8GB.