ImportError: cannot import name 'CLIPVisionModelWithProjection' from 'transformers'

[mothormothormothor]

Hey there,

I get this error since today. I changed nothing until the last time I used the notebook. Does somebody have a solution for this ?
I'm running Colab on a A100.

thanks
xoxo

╭─────────────────────────────── Traceback (most recent call last) ────────────────────────────────╮
│ /content/stable-dreamfusion/main.py:138 in │
│ │
│ 135 │ │ # scheduler = lambda optimizer: optim.lr_scheduler.OneCycleLR(optimizer, max_lr= │
│ 136 │ │ │
│ 137 │ │ if opt.guidance == 'stable-diffusion': │
│ ❱ 138 │ │ │ from nerf.sd import StableDiffusion │
│ 139 │ │ │ guidance = StableDiffusion(device) │
│ 140 │ │ elif opt.guidance == 'clip': │
│ 141 │ │ │ from nerf.clip import CLIP │
│ │
│ /content/stable-dreamfusion/nerf/sd.py:2 in │
│ │
│ 1 from transformers import CLIPTextModel, CLIPTokenizer, logging │
│ ❱ 2 from diffusers import AutoencoderKL, UNet2DConditionModel, PNDMScheduler │
│ 3 │
│ 4 # suppress partial model loading warning │
│ 5 logging.set_verbosity_error() │
│ │
│ /usr/local/lib/python3.7/dist-packages/diffusers/init.py:32 in │
│ │
│ 29 │ │ get_scheduler, │
│ 30 │ ) │
│ 31 │ from .pipeline_utils import DiffusionPipeline │
│ ❱ 32 │ from .pipelines import ( │
│ 33 │ │ DanceDiffusionPipeline, │
│ 34 │ │ DDIMPipeline, │
│ 35 │ │ DDPMPipeline, │
│ │
│ /usr/local/lib/python3.7/dist-packages/diffusers/pipelines/init.py:18 in │
│ │
│ 15 │ from ..utils.dummy_pt_objects import * # noqa F403 │
│ 16 │
│ 17 if is_torch_available() and is_transformers_available(): │
│ ❱ 18 │ from .alt_diffusion import AltDiffusionImg2ImgPipeline, AltDiffusionPipeline │
│ 19 │ from .latent_diffusion import LDMTextToImagePipeline │
│ 20 │ from .stable_diffusion import ( │
│ 21 │ │ CycleDiffusionPipeline, │
│ │
│ /usr/local/lib/python3.7/dist-packages/diffusers/pipelines/alt_diffusion/init.py:33 in │
│ │
│ │
│ 30 │
│ 31 if is_transformers_available() and is_torch_available(): │
│ 32 │ from .modeling_roberta_series import RobertaSeriesModelWithTransformation │
│ ❱ 33 │ from .pipeline_alt_diffusion import AltDiffusionPipeline │
│ 34 │ from .pipeline_alt_diffusion_img2img import AltDiffusionImg2ImgPipeline │
│ 35 │
│ │
│ /usr/local/lib/python3.7/dist-packages/diffusers/pipelines/alt_diffusion/pipeline_alt_diffusion. │
│ py:35 in │
│ │
│ 32 │ PNDMScheduler, │
│ 33 ) │
│ 34 from ...utils import deprecate, logging │
│ ❱ 35 from ..stable_diffusion.safety_checker import StableDiffusionSafetyChecker │
│ 36 from . import AltDiffusionPipelineOutput, RobertaSeriesModelWithTransformation │
│ 37 │
│ 38 │
│ │
│ /usr/local/lib/python3.7/dist-packages/diffusers/pipelines/stable_diffusion/init.py:33 in │
│ │
│ │
│ 30 if is_transformers_available() and is_torch_available(): │
│ 31 │ from .pipeline_cycle_diffusion import CycleDiffusionPipeline │
│ 32 │ from .pipeline_stable_diffusion import StableDiffusionPipeline │
│ ❱ 33 │ from .pipeline_stable_diffusion_image_variation import StableDiffusionImageVariation │
│ 34 │ from .pipeline_stable_diffusion_img2img import StableDiffusionImg2ImgPipeline │
│ 35 │ from .pipeline_stable_diffusion_inpaint import StableDiffusionInpaintPipeline │
│ 36 │ from .pipeline_stable_diffusion_inpaint_legacy import StableDiffusionInpaintPipeline │
│ │
│ /usr/local/lib/python3.7/dist-packages/diffusers/pipelines/stable_diffusion/pipeline_stable_diff │
│ usion_image_variation.py:22 in │
│ │
│ 19 │
│ 20 import PIL │
│ 21 from diffusers.utils import is_accelerate_available │
│ ❱ 22 from transformers import CLIPFeatureExtractor, CLIPVisionModelWithProjection │
│ 23 │
│ 24 from ...models import AutoencoderKL, UNet2DConditionModel │
│ 25 from ...pipeline_utils import DiffusionPipeline │
╰──────────────────────────────────────────────────────────────────────────────────────────────────╯
ImportError: cannot import name 'CLIPVisionModelWithProjection' from 'transformers'
(/usr/local/lib/python3.7/dist-packages/transformers/init.py)

[gianluigidalessandro]

I am having the same problem.

[XinandYu]

huggingface/diffusers#1377

This might be helpful

[anton-l]

Fixed in a patch release 0.8.1: https://github.com/huggingface/diffusers/releases/tag/v0.8.1
⚠️ Please upgrade with pip install --upgrade diffusers or pip install diffusers==0.8.1

[mothormothormothor]

pip install diffusers==0.8.1

this one worked! thank you very much :-)

[saipheblue]

• pip install diffusers==0.8.1

worked for me as well ! Thank you