Cannot import ClipProcessor and other bugs #3028
Description
Describe the bug
There is a constant import error:
ImportError: cannot import name 'CLIPImageProcessor' from 'transformers' (D:\Python\lib\site-packages\transformers\__init__.py)
Somehow managed to fix it, but still another problem: no matter what I write, NSFW content is always determined
When disabling safety checker, black photo is still saved
Help pls
Reproduction
from diffusers import StableDiffusionPipeline
import torch
model_id = "runwayml/stable-diffusion-v1-5"
pipe = StableDiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.float16)
pipe = pipe.to("cuda")
prompt = "a photo of an astronaut riding a horse on mars"
image = pipe(prompt).images[0]
image.save("astronaut_rides_horse.png")
Logs
╭─────────────────────────────── Traceback (most recent call last) ────────────────────────────────╮
│ │
│ D:\PycharmProjects\gpt\txtvideo2.py:1 in <module> │
│ │
│ ❱ 1 from diffusers import StableDiffusionPipeline │
│ 2 import torch │
│ 3 │
│ 4 model_id = "runwayml/stable-diffusion-v1-5" │
│ D:\Python\lib\site-packages\diffusers\__init__.py:58 in <module> │
│ │
│ 55 │ │ get_polynomial_decay_schedule_with_warmup, │
│ 56 │ │ get_scheduler, │
│ 57 │ ) │
│ ❱ 58 │ from .pipelines import ( │
│ 59 │ │ AudioPipelineOutput, │
│ 60 │ │ DanceDiffusionPipeline, │
│ 61 │ │ DDIMPipeline, │
│ │
│ D:\Python\lib\site-packages\diffusers\pipelines\__init__.py:45 in <module> │
│ │
│ 42 except OptionalDependencyNotAvailable: │
│ 43 │ from ..utils.dummy_torch_and_transformers_objects import * # noqa F403 │
│ 44 else: │
│ ❱ 45 │ from .alt_diffusion import AltDiffusionImg2ImgPipeline, AltDiffusionPipeline │
│ 46 │ from .audioldm import AudioLDMPipeline │
│ 47 │ from .latent_diffusion import LDMTextToImagePipeline │
│ 48 │ from .paint_by_example import PaintByExamplePipeline │
│ │
│ D:\Python\lib\site-packages\diffusers\pipelines\alt_diffusion\__init__.py:32 in <module> │
│ │
│ 29 │
│ 30 if is_transformers_available() and is_torch_available(): │
│ 31 │ from .modeling_roberta_series import RobertaSeriesModelWithTransformation │
│ ❱ 32 │ from .pipeline_alt_diffusion import AltDiffusionPipeline │
│ 33 │ from .pipeline_alt_diffusion_img2img import AltDiffusionImg2ImgPipeline │
│ 34 │
│ │
│ D:\Python\lib\site-packages\diffusers\pipelines\alt_diffusion\pipeline_alt_diffusion.py:20 in │
│ <module> │
│ │
│ 17 │
│ 18 import torch │
│ 19 from packaging import version │
│ ❱ 20 from transformers import CLIPImageProcessor, XLMRobertaTokenizer │
│ 21 │
│ 22 from diffusers.utils import is_accelerate_available, is_accelerate_version │
│ 23 │
╰──────────────────────────────────────────────────────────────────────────────────────────────────╯
ImportError: cannot import name 'CLIPImageProcessor' from 'transformers'
(D:\Python\lib\site-packages\transformers\__init__.py)System Info
Diffusers version: 0.15.0.dev0
transformers version: 4.23.1
torch version: 2.0.0+cu117
Python version: 3.10.6
OS: Windows 10 x64 Pro
VideoCard: NVIDIA GTX 1660 SUPER TI
Processor: Intel pentium gold 3.7 GHz
Locale: RU