Running Spacenet Vegas Semantic Segmentation Example

[yashgharat]

🐛 Bug

So when I follow the instructions and have a local copy of the spacenet vegas folder, I get the error:

root@cca5d668d324:/opt/src# rastervision run local code/spacenet_vegas.py -a raw_uri $RAW_URI -a root_uri $ROOT_URI
Traceback (most recent call last):
  File "/opt/conda/lib/python3.7/runpy.py", line 193, in _run_module_as_main
    "__main__", mod_spec)
  File "/opt/conda/lib/python3.7/runpy.py", line 85, in _run_code
    exec(code, run_globals)
  File "/opt/src/rastervision_pipeline/rastervision/pipeline/cli.py", line 250, in <module>
    _main()
  File "/opt/src/rastervision_pipeline/rastervision/pipeline/cli.py", line 246, in _main
    main()
  File "/opt/conda/lib/python3.7/site-packages/click/core.py", line 1130, in __call__
    return self.main(*args, **kwargs)
  File "/opt/conda/lib/python3.7/site-packages/click/core.py", line 1055, in main
    rv = self.invoke(ctx)
  File "/opt/conda/lib/python3.7/site-packages/click/core.py", line 1657, in invoke
    return _process_result(sub_ctx.command.invoke(sub_ctx))
  File "/opt/conda/lib/python3.7/site-packages/click/core.py", line 1404, in invoke
    return ctx.invoke(self.callback, **ctx.params)
  File "/opt/conda/lib/python3.7/site-packages/click/core.py", line 760, in invoke
    return __callback(*args, **kwargs)
  File "/opt/src/rastervision_pipeline/rastervision/pipeline/cli.py", line 170, in run
    cfgs = get_configs(cfg_module, runner, args)
  File "/opt/src/rastervision_pipeline/rastervision/pipeline/cli.py", line 74, in get_configs
    cfgs = _get_configs(runner, **args)
  File "code/spacenet_vegas.py", line 229, in get_config
    chip_options=chip_options
  File "pydantic/main.py", line 331, in pydantic.main.BaseModel.__init__
pydantic.error_wrappers.ValidationError: 1 validation error for SemanticSegmentationConfig
img_format
  extra fields not permitted (type=value_error.extra)

This references the img_format option on line 227 although it mentions line 228. However, I notice that the tiny_spacenet example and the actual SemanticSegmentationConfig Class say those two params are optional. so when removing it to test, I end with another error.

root@cca5d668d324:/opt/src# rastervision run local code/spacenet_vegas.py -a raw_uri $RAW_URI -a root_uri $ROOT_URI
Traceback (most recent call last):
  File "/opt/conda/lib/python3.7/runpy.py", line 193, in _run_module_as_main
    "__main__", mod_spec)
  File "/opt/conda/lib/python3.7/runpy.py", line 85, in _run_code
    exec(code, run_globals)
  File "/opt/src/rastervision_pipeline/rastervision/pipeline/cli.py", line 250, in <module>
    _main()
  File "/opt/src/rastervision_pipeline/rastervision/pipeline/cli.py", line 246, in _main
    main()
  File "/opt/conda/lib/python3.7/site-packages/click/core.py", line 1130, in __call__
    return self.main(*args, **kwargs)
  File "/opt/conda/lib/python3.7/site-packages/click/core.py", line 1055, in main
    rv = self.invoke(ctx)
  File "/opt/conda/lib/python3.7/site-packages/click/core.py", line 1657, in invoke
    return _process_result(sub_ctx.command.invoke(sub_ctx))
  File "/opt/conda/lib/python3.7/site-packages/click/core.py", line 1404, in invoke
    return ctx.invoke(self.callback, **ctx.params)
  File "/opt/conda/lib/python3.7/site-packages/click/core.py", line 760, in invoke
    return __callback(*args, **kwargs)
  File "/opt/src/rastervision_pipeline/rastervision/pipeline/cli.py", line 175, in run
    pipeline_run_name)
  File "/opt/src/rastervision_pipeline/rastervision/pipeline/cli.py", line 117, in _run_pipeline
    build_config(cfg.dict())
  File "/opt/src/rastervision_pipeline/rastervision/pipeline/config.py", line 163, in build_config
    new_x[k] = build_config(v)
  File "/opt/src/rastervision_pipeline/rastervision/pipeline/config.py", line 163, in build_config
    new_x[k] = build_config(v)
  File "/opt/src/rastervision_pipeline/rastervision/pipeline/config.py", line 167, in build_config
    new_x = config_cls(**new_x)
  File "pydantic/main.py", line 331, in pydantic.main.BaseModel.__init__
pydantic.error_wrappers.ValidationError: 1 validation error for SemanticSegmentationGeoDataConfig
__root__
  len(class_colors) != len(class_names) (type=value_error.config)

I also agree that these above errors are from pulling the master directly. However, when I run the 0.13 branch, I still get an error that says

Traceback (most recent call last):
  File "/opt/conda/lib/python3.6/runpy.py", line 183, in _run_module_as_main
    mod_name, mod_spec, code = _get_module_details(mod_name, _Error)
  File "/opt/conda/lib/python3.6/runpy.py", line 109, in _get_module_details
    __import__(pkg_name)
  File "/opt/src/rastervision_pipeline/rastervision/pipeline/__init__.py", line 19, in <module>
    registry.load_plugins()
  File "/opt/src/rastervision_pipeline/rastervision/pipeline/registry.py", line 238, in load_plugins
    for finder, name, ispkg in iter_namespace(rastervision)
  File "/opt/src/rastervision_pipeline/rastervision/pipeline/registry.py", line 238, in <dictcomp>
    for finder, name, ispkg in iter_namespace(rastervision)
  File "/opt/conda/lib/python3.6/importlib/__init__.py", line 126, in import_module
    return _bootstrap._gcd_import(name[level:], package, level)
  File "/opt/src/rastervision_pytorch_backend/rastervision/pytorch_backend/__init__.py", line 4, in <module>
    from rastervision.pytorch_backend.pytorch_chip_classification_config import *
  File "/opt/src/rastervision_pytorch_backend/rastervision/pytorch_backend/pytorch_chip_classification_config.py", line 2, in <module>
    from rastervision.pytorch_backend.pytorch_learner_backend_config import (
  File "/opt/src/rastervision_pytorch_backend/rastervision/pytorch_backend/pytorch_learner_backend_config.py", line 6, in <module>
    from rastervision.pytorch_learner.learner_config import (
  File "/opt/src/rastervision_pytorch_learner/rastervision/pytorch_learner/__init__.py", line 4, in <module>
    from rastervision.pytorch_learner.learner_config import *
  File "/opt/src/rastervision_pytorch_learner/rastervision/pytorch_learner/learner_config.py", line 11, in <module>
    import albumentations as A
  File "/opt/conda/lib/python3.6/site-packages/albumentations/__init__.py", line 5, in <module>
    from .core.composition import *
  File "/opt/conda/lib/python3.6/site-packages/albumentations/core/composition.py", line 8, in <module>
    from albumentations.augmentations.keypoints_utils import KeypointsProcessor
  File "/opt/conda/lib/python3.6/site-packages/albumentations/augmentations/__init__.py", line 4, in <module>
    from .functional import *
  File "/opt/conda/lib/python3.6/site-packages/albumentations/augmentations/functional.py", line 7, in <module>
    import cv2
  File "/opt/conda/lib/python3.6/site-packages/cv2/__init__.py", line 181, in <module>
    bootstrap()
  File "/opt/conda/lib/python3.6/site-packages/cv2/__init__.py", line 175, in bootstrap
    if __load_extra_py_code_for_module("cv2", submodule, DEBUG):
  File "/opt/conda/lib/python3.6/site-packages/cv2/__init__.py", line 28, in __load_extra_py_code_for_module
    py_module = importlib.import_module(module_name)
  File "/opt/conda/lib/python3.6/importlib/__init__.py", line 126, in import_module
    return _bootstrap._gcd_import(name[level:], package, level)
  File "/opt/conda/lib/python3.6/site-packages/cv2/mat_wrapper/__init__.py", line 33, in <module>
    cv._registerMatType(Mat)
AttributeError: module 'cv2' has no attribute '_registerMatType'

To Reproduce

Steps to reproduce the behavior:

1. Follow the setup instructions
2. Mount Spacenet Vegas data locally
3. Run the rastervision model with rastervision run local code/spacenet_vegas.py -a raw_uri $RAW_URI -a root_uri $ROOT_URI -a test FALSE

The config can be found in the experiment
but just in case:

   return SemanticSegmentationConfig(
       root_uri=root_uri,
       dataset=scene_dataset,
       backend=backend,
       train_chip_sz=chip_sz,
       predict_chip_sz=chip_sz,
       #img_format='npy',
       chip_options=chip_options
   )

Expected behavior

Environment

Running Raster Vision directly in Windows is not supported, and we recommend that you run it from within a Docker container.

• How you installed and are running Raster Vision (pip install on local vs. inside Docker image):
  • I'm using the docker/build and docker/run --gpu from the Github and mounting my local data with it
• Raster Vision version or commit:
  • 0.13, just pulled it last week and followed the setup instructions
• OS (e.g., Linux):
  • Ubuntu 22.04
• Python version:
  • 3.6.9
• CUDA/cuDNN version if running on GPU:
  • 10.2 because of the docker

Any direction of the source of the errors would be helpful.

[yashgharat]

So the master branch does not work.
But the 0.13 needed an openCV update. I installed with pip install opencv-python-headless==4.5.4.60

Not sure to close this since the first two error messages are still valid. Feel free to close otherwise.

[AdeelH]

Hi, thanks for the report!

Yes, some examples are not yet updated to be compatible with the master branch. I would recommend using the 0.13 branch for now.

I would keep this PR open because this is something we definitely need to fix.

[AdeelH]

Did installing opencv via pip fix the problem?

[yashgharat]

@AdeelH for the 0.13 branch yes installing opencv fixed the line. I added this line to the Dockerfile.
RUN pip install opencv-python-headless==4.5.4.60
that solved the error for
AttributeError: module 'cv2' has no attribute '_registerMatType'

[AdeelH]

BTW, this same opencv error was fixed on master earlier this year in this PR: #1323