feat(docker): blackwell GPU compatibility

[amadeuszsz]

Summary

Update core dependencies.

Note

Test performed

[KSeangTan]

LGTM!

[vividf]

@amadeuszsz
Seems that the mmdet version is not compatible with mmcv=2.2.0

Traceback (most recent call last): File "/workspace/projects/CalibrationStatusClassification/deploy/main.py", line 38, in <module> from autoware_ml.calibration_classification.datasets.transforms.calibration_classification_transform import ( File "/workspace/autoware_ml/__init__.py", line 3, in <module> import mmdet File "/opt/conda/lib/python3.11/site-packages/mmdet/__init__.py", line 16, in <module> assert (mmcv_version >= digit_version(mmcv_minimum_version) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ AssertionError: MMCV==2.2.0 is used but incompatible. Please install mmcv>=2.0.0rc4, <2.2.0. root@66c4ae6d9b78:/workspace# ^C root@66c4ae6d9b78:/workspace# ^C root@66c4ae6d9b78:/workspace# python projects/CalibrationStatusClassification/deploy/main.py projects/CalibrationStatusClassification/configs/deploy/resnet18_5ch.py projects/CalibrationStatusClassification/configs/t4dataset/resnet18_5ch_1xb16-50e_j6gen2.py Traceback (most recent call last): File "/workspace/projects/CalibrationStatusClassification/deploy/main.py", line 38, in <module> from autoware_ml.calibration_classification.datasets.transforms.calibration_classification_transform import ( File "/workspace/autoware_ml/__init__.py", line 3, in <module> import mmdet File "/opt/conda/lib/python3.11/site-packages/mmdet/__init__.py", line 16, in <module> assert (mmcv_version >= digit_version(mmcv_minimum_version) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ AssertionError: MMCV==2.2.0 is used but incompatible. Please install mmcv>=2.0.0rc4, <2.2.0.

[amadeuszsz]

@amadeuszsz Seems that the mmdet version is not compatible with mmcv=2.2.0

Traceback (most recent call last): File "/workspace/projects/CalibrationStatusClassification/deploy/main.py", line 38, in <module> from autoware_ml.calibration_classification.datasets.transforms.calibration_classification_transform import ( File "/workspace/autoware_ml/__init__.py", line 3, in <module> import mmdet File "/opt/conda/lib/python3.11/site-packages/mmdet/__init__.py", line 16, in <module> assert (mmcv_version >= digit_version(mmcv_minimum_version) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ AssertionError: MMCV==2.2.0 is used but incompatible. Please install mmcv>=2.0.0rc4, <2.2.0. root@66c4ae6d9b78:/workspace# ^C root@66c4ae6d9b78:/workspace# ^C root@66c4ae6d9b78:/workspace# python projects/CalibrationStatusClassification/deploy/main.py projects/CalibrationStatusClassification/configs/deploy/resnet18_5ch.py projects/CalibrationStatusClassification/configs/t4dataset/resnet18_5ch_1xb16-50e_j6gen2.py Traceback (most recent call last): File "/workspace/projects/CalibrationStatusClassification/deploy/main.py", line 38, in <module> from autoware_ml.calibration_classification.datasets.transforms.calibration_classification_transform import ( File "/workspace/autoware_ml/__init__.py", line 3, in <module> import mmdet File "/opt/conda/lib/python3.11/site-packages/mmdet/__init__.py", line 16, in <module> assert (mmcv_version >= digit_version(mmcv_minimum_version) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ AssertionError: MMCV==2.2.0 is used but incompatible. Please install mmcv>=2.0.0rc4, <2.2.0.

@vividf
Good catch! Even though we can probably patch it, I think we don't need so far any feature from 2.2.0. Addressed in 5b10dc5

[vividf]

  File "/opt/conda/lib/python3.11/site-packages/mmengine/runner/runner.py", line 2127, in load_checkpoint
    checkpoint = _load_checkpoint(filename, map_location=map_location)
                 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/opt/conda/lib/python3.11/site-packages/mmengine/runner/checkpoint.py", line 548, in _load_checkpoint
    return CheckpointLoader.load_checkpoint(filename, map_location, logger)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/opt/conda/lib/python3.11/site-packages/mmengine/runner/checkpoint.py", line 330, in load_checkpoint
    return checkpoint_loader(filename, map_location)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/opt/conda/lib/python3.11/site-packages/mmengine/runner/checkpoint.py", line 347, in load_from_local
    checkpoint = torch.load(filename, map_location=map_location)
                 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/opt/conda/lib/python3.11/site-packages/torch/serialization.py", line 1529, in load
    raise pickle.UnpicklingError(_get_wo_message(str(e))) from None
_pickle.UnpicklingError: Weights only load failed. This file can still be loaded, to do so you have two options, do those steps only if you trust the source of the checkpoint. 
	(1) In PyTorch 2.6, we changed the default value of the `weights_only` argument in `torch.load` from `False` to `True`. Re-running `torch.load` with `weights_only` set to `False` will likely succeed, but it can result in arbitrary code execution. Do it only if you got the file from a trusted source.
	(2) Alternatively, to load with `weights_only=True` please check the recommended steps in the following error message.
	WeightsUnpickler error: Unsupported global: GLOBAL mmengine.logging.history_buffer.HistoryBuffer was not an allowed global by default. Please use `torch.serialization.add_safe_globals([mmengine.logging.history_buffer.HistoryBuffer])` or the `torch.serialization.safe_globals([mmengine.logging.history_buffer.HistoryBuffer])` context manager to allowlist this global if you trust this class/function.

Check the documentation of torch.load to learn more about types accepted by default with weights_only https://pytorch.org/docs/stable/generated/torch.load.html.

I think some additional work needs to be done (trying now)

[amadeuszsz]

  File "/opt/conda/lib/python3.11/site-packages/mmengine/runner/runner.py", line 2127, in load_checkpoint
    checkpoint = _load_checkpoint(filename, map_location=map_location)
                 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/opt/conda/lib/python3.11/site-packages/mmengine/runner/checkpoint.py", line 548, in _load_checkpoint
    return CheckpointLoader.load_checkpoint(filename, map_location, logger)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/opt/conda/lib/python3.11/site-packages/mmengine/runner/checkpoint.py", line 330, in load_checkpoint
    return checkpoint_loader(filename, map_location)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/opt/conda/lib/python3.11/site-packages/mmengine/runner/checkpoint.py", line 347, in load_from_local
    checkpoint = torch.load(filename, map_location=map_location)
                 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/opt/conda/lib/python3.11/site-packages/torch/serialization.py", line 1529, in load
    raise pickle.UnpicklingError(_get_wo_message(str(e))) from None
_pickle.UnpicklingError: Weights only load failed. This file can still be loaded, to do so you have two options, do those steps only if you trust the source of the checkpoint. 
	(1) In PyTorch 2.6, we changed the default value of the `weights_only` argument in `torch.load` from `False` to `True`. Re-running `torch.load` with `weights_only` set to `False` will likely succeed, but it can result in arbitrary code execution. Do it only if you got the file from a trusted source.
	(2) Alternatively, to load with `weights_only=True` please check the recommended steps in the following error message.
	WeightsUnpickler error: Unsupported global: GLOBAL mmengine.logging.history_buffer.HistoryBuffer was not an allowed global by default. Please use `torch.serialization.add_safe_globals([mmengine.logging.history_buffer.HistoryBuffer])` or the `torch.serialization.safe_globals([mmengine.logging.history_buffer.HistoryBuffer])` context manager to allowlist this global if you trust this class/function.

Check the documentation of torch.load to learn more about types accepted by default with weights_only https://pytorch.org/docs/stable/generated/torch.load.html.

If we used PyTorch 2.6,the mmengine checkpoint loading code needs to be fixed (or maybe updated? checking right now)

@vividf
Right, already noted same issue in PTv3 PR. Looks like we need to patch openmmlab packages. Thanks! I will also update AWML projects which use torch.load.