stable-diffusion-deploy
                                
                                 stable-diffusion-deploy copied to clipboard
                                
                                    stable-diffusion-deploy copied to clipboard
                            
                            
                            
                        Cannot import name 'Image' from 'PIL' (unknown location)
🐛 Bug
I am trying to create a Lightning BYOC cluster
To Reproduce
Steps to reproduce the behavior:
- Run lightning create cluster <cluster-name> --role-arn <ARN> --external-id <EXTERNAL-ID>
Traceback (most recent call last):
  File "/usr/local/bin/lightning", line 5, in <module>
    from lightning.app.cli.lightning_cli import main
  File "/usr/local/lib/python3.9/site-packages/lightning/__init__.py", line 32, in <module>
    from lightning.app import storage  # noqa: E402
  File "/usr/local/lib/python3.9/site-packages/lightning/app/__init__.py", line 25, in <module>
    from lightning.app import components  # noqa: E402, F401
  File "/usr/local/lib/python3.9/site-packages/lightning/app/components/__init__.py", line 15, in <module>
    from lightning.app.components.serve.serve import ModelInferenceAPI
  File "/usr/local/lib/python3.9/site-packages/lightning/app/components/serve/serve.py", line 28, in <module>
    from lightning.app.components.serve.types import _DESERIALIZER, _SERIALIZER
  File "/usr/local/lib/python3.9/site-packages/lightning/app/components/serve/types/__init__.py", line 1, in <module>
    from lightning.app.components.serve.types.image import Image
  File "/usr/local/lib/python3.9/site-packages/lightning/app/components/serve/types/image.py", line 25, in <module>
    from PIL import Image as PILImage
ImportError: cannot import name 'Image' from 'PIL' (unknown location)
Code sample
Expected behavior
I'm expecting the command to work.
Environment
- PyTorch Version (e.g., 1.0):
- OS (e.g., Linux): macOS
- How you installed PyTorch (conda,pip, source): virtualenv
- Build command you used (if compiling from source):
- Python version: 3.9
- CUDA/cuDNN version:
- GPU models and configuration:
- Any other relevant information:
Additional context
I tried the following without success:
pip uninstall Pillow
pip install Pillow
I also tried:
pip install Pillow==8.3.1
It looks like your environment is broken or has some inconsistencies. Could you try installing lightning in a new environment (you don't need to install Pillow explicitly - just pip install -U lightning would do) and try again?
@hhsecond same issue. I tried pip, conda, virtualenv, and virtualenvwrapper. I also tried pip install -U lightning but the issue persists. Could this be a dependency issue, examplePIL? Pillow was a PIL fork.