trying to get Object detection to work

[georgeper2001]

Initially the scripts were written for chip detection. I found an example of object detect and made changes. I have attached my python script file chevron_pj_object2.zip. I use the command:

rastervision run local chevron_pj_object2.py -a root_uri /tmp/hello -a phase3_1 chevron_pool_NM.json -a phase3_2 base_pool.json -a phase7 chevron_pool_KERN.json -a phase8 base_abnormal.json -a phase9 chevron_pool_SM2.json -a batch_size 12 -a epochs 32 -a arch resnet-50 analyze chip train

It goes through the analyze process fine. When it gets to the train process it gets this error:

2024-03-18 18:17:07:rastervision.pytorch_learner.learner: INFO - Devices:
index, name, driver_version, memory.total [MiB], memory.used [MiB], memory.free [MiB]
0, NVIDIA GeForce RTX 3080, 525.147.05, 10240 MiB, 169 MiB, 9831 MiB
2024-03-18 18:17:07:rastervision.pytorch_learner.learner: INFO - PyTorch version: 1.7.1
2024-03-18 18:17:07:rastervision.pytorch_learner.learner: INFO - CUDA available: True
2024-03-18 18:17:07:rastervision.pytorch_learner.learner: INFO - CUDA version: 11.0
2024-03-18 18:17:07:rastervision.pytorch_learner.learner: INFO - CUDNN version: 8005
2024-03-18 18:17:07:rastervision.pytorch_learner.learner: INFO - Number of CUDA devices: 1
2024-03-18 18:17:07:rastervision.pytorch_learner.learner: INFO - Active CUDA Device: GPU 0
Traceback (most recent call last):
  File "/opt/conda/lib/python3.6/runpy.py", line 193, in _run_module_as_main
    "__main__", mod_spec)
  File "/opt/conda/lib/python3.6/runpy.py", line 85, in _run_code
    exec(code, run_globals)
  File "/opt/src/rastervision_pipeline/rastervision/pipeline/cli.py", line 248, in <module>
    main()
  File "/opt/conda/lib/python3.6/site-packages/click/core.py", line 722, in __call__
    return self.main(*args, **kwargs)
  File "/opt/conda/lib/python3.6/site-packages/click/core.py", line 697, in main
    rv = self.invoke(ctx)
  File "/opt/conda/lib/python3.6/site-packages/click/core.py", line 1066, in invoke
    return _process_result(sub_ctx.command.invoke(sub_ctx))
  File "/opt/conda/lib/python3.6/site-packages/click/core.py", line 895, in invoke
    return ctx.invoke(self.callback, **ctx.params)
  File "/opt/conda/lib/python3.6/site-packages/click/core.py", line 535, in invoke
    return callback(*args, **kwargs)
  File "/opt/src/rastervision_pipeline/rastervision/pipeline/cli.py", line 240, in run_command
    runner=runner)
  File "/opt/src/rastervision_pipeline/rastervision/pipeline/cli.py", line 217, in _run_command
    command_fn()
  File "/opt/src/rastervision_core/rastervision/core/rv_pipeline/rv_pipeline.py", line 134, in train
    backend.train(source_bundle_uri=self.config.source_bundle_uri)
  File "/opt/src/rastervision_pytorch_backend/rastervision/pytorch_backend/pytorch_learner_backend.py", line 74, in train
    learner = self.learner_cfg.build(self.tmp_dir, training=True)
  File "/opt/src/rastervision_pytorch_learner/rastervision/pytorch_learner/object_detection_learner_config.py", line 155, in build
    training=training)
  File "/opt/src/rastervision_pytorch_learner/rastervision/pytorch_learner/learner.py", line 155, in __init__
    self.setup_model(model_def_path=model_def_path)
  File "/opt/src/rastervision_pytorch_learner/rastervision/pytorch_learner/learner.py", line 273, in setup_model
    self.model = self.build_model()
  File "/opt/src/rastervision_pytorch_learner/rastervision/pytorch_learner/object_detection_learner.py", line 28, in build_model
    pretrained=pretrained)
  File "/opt/src/rastervision_pytorch_learner/rastervision/pytorch_learner/object_detection_utils.py", line 404, in __init__
    backbone = resnet_fpn_backbone(backbone_arch, pretrained)
  File "/opt/src/rastervision_pytorch_learner/rastervision/pytorch_learner/object_detection_utils.py", line 337, in resnet_fpn_backbone
    pretrained=pretrained, norm_layer=misc_nn_ops.FrozenBatchNorm2d)
  File "/opt/conda/lib/python3.6/site-packages/torchvision/models/resnet.py", line 186, in resnet50
    model = ResNet(Bottleneck, [3, 4, 6, 3], **kwargs)
TypeError: __init__() got an unexpected keyword argument 'norm_layer'
/tmp/hello/Makefile:12: recipe for target '2' failed
make: *** [2] Error 1

I need help on what to do, any help is appreciated.
attached file: chevron_pj_object2.zip

[AdeelH]

This looks like an error from an older Raster Vision version. I recommend using a newer docker image (https://quay.io/repository/azavea/raster-vision?tab=tags).

Also, please make sure to remove any potentially sensitive/confidential information when sharing code.

[georgeper2001]

I am using the newer docker image, but it is using a different version of pystac. I am getting the error: "
ImportError: cannot import name 'STAC_IO' from 'pystac' (/opt/conda/lib/python3.10/site-packages/pystac/init.py)
"
The error is coming from this line: 'from pystac import STAC_IO, Catalog'
can I get it to use an older version of pystac? I am guessing that pystac now has pystac.stac_io; which also uses read_text and write_text instead of read_text_method and write_text_metod.

[georgeper2001]

Or how would I fix the error?

[AdeelH]

You will probably need to update the STAC-related parts of the code to be compatible with the newer pystac version.

I would also recommend trying out the read_stac() function in Raster Vision. I suspect you might be able to replace a significant portion of the existing STAC-reading code with a single call to this function.

[georgeper2001]

Just FYI. I needed at least pytorch-1.10.2 to make the script work.