TensorRT fails with calibration

 cm run script --tags=run-mlperf,inference,_r4.1-dev,_all-scenarios \
                --model=retinanet \
                --implementation=nvidia \
                --framework=tensorrt \
                --category=datacenter \
                --server_target_qps=50 \
                --execution_mode=valid \
                --device=cuda \
                        --repro \
                --quiet \

I ran into an issue with calibration. It was the same with different Inference version (v5.0)
How can I solve this issue?

make calibrate RUN_ARGS=' --benchmarks=retinanet --scenarios=server  --test_mode=PerformanceOnly  --server_target_qps=50 --gpu_copy_streams=2 --gpu_inference_streams=2 --gpu_batch_size=8 --use_deque_limit --no_audit_verify  '
[2025-06-12 15:13:32,657 main.py:229 INFO] Detected system ID: KnownSystem.Nvidia_5e0adc8add39
[2025-06-12 15:13:32,814 calibrate.py:45 INFO] Generating calibration cache for Benchmark "retinanet"
[06/12/2025-15:13:34] [TRT] [I] [MemUsageChange] Init CUDA: CPU +2, GPU +0, now: CPU 364, GPU 924 (MiB)
[06/12/2025-15:13:41] [TRT] [I] [MemUsageChange] Init builder kernel library: CPU +4417, GPU +1160, now: CPU 4917, GPU 2084 (MiB)
[2025-06-12 15:13:42,059 retinanet_graphsurgeon.py:159 INFO] Renaming layers...
[2025-06-12 15:13:42,060 retinanet_graphsurgeon.py:237 INFO] Renamed 225 layers.
[2025-06-12 15:13:42,060 retinanet_graphsurgeon.py:243 INFO] Renaming tensors to match layer names
[2025-06-12 15:13:42,063 retinanet_graphsurgeon.py:264 INFO] Adding NMS layer nmsopt to the graph...
[06/12/2025-15:13:42] [TRT] [I] No checker registered for op: NMS_OPT_TRT. Attempting to check as plugin.
[06/12/2025-15:13:42] [TRT] [I] No importer registered for op: NMS_OPT_TRT. Attempting to import as plugin.
[06/12/2025-15:13:42] [TRT] [I] Searching for plugin: NMS_OPT_TRT, plugin_version: 2, plugin_namespace: 
[06/12/2025-15:13:42] [TRT] [W] builtin_op_importers.cpp:5677: Attribute permuteBeforeReshape not found in plugin node! Ensure that the plugin creator has a default value defined or the engine may fail to build.
[06/12/2025-15:13:42] [TRT] [W] builtin_op_importers.cpp:5677: Attribute concatInputs not found in plugin node! Ensure that the plugin creator has a default value defined or the engine may fail to build.
[06/12/2025-15:13:42] [TRT] [I] Successfully created plugin: NMS_OPT_TRT
/usr/local/lib/python3.8/dist-packages/torch/functional.py:504: UserWarning: torch.meshgrid: in an upcoming release, it will be required to pass the indexing argument. (Triggered internally at /home/cmuser/CM/repos/local/cache/ac4a8632ea8a437d/pytorch/aten/src/ATen/native/TensorShape.cpp:3516.)
  return _VF.meshgrid(tensors, **kwargs)  # type: ignore[attr-defined]
Traceback (most recent call last):
  File "/root/CM/repos/local/cache/db6b5d49a322435e/repo/closed/NVIDIA/code/actionhandler/base.py", line 78, in run
    success = self.handle()
  File "/root/CM/repos/local/cache/db6b5d49a322435e/repo/closed/NVIDIA/code/actionhandler/calibrate.py", line 62, in handle
    b.calibrate()
  File "/usr/local/lib/python3.8/dist-packages/nvmitten/nvidia/builder.py", line 594, in calibrate
    self.mitten_builder.run(self.legacy_scratch, None)
  File "/usr/local/lib/python3.8/dist-packages/nvmitten/debug/debug_manager.py", line 258, in _wrapper
    raise exc_info[1]
  File "/usr/local/lib/python3.8/dist-packages/nvmitten/debug/debug_manager.py", line 245, in _wrapper
    retval = obj(*args, **kwargs)
  File "/root/CM/repos/local/cache/db6b5d49a322435e/repo/closed/NVIDIA/code/retinanet/tensorrt/Retinanet.py", line 379, in run
    network = self.create_network(self.builder, subnetwork_name=subnet_name)
  File "/usr/local/lib/python3.8/dist-packages/nvmitten/debug/debug_manager.py", line 258, in _wrapper
    raise exc_info[1]
  File "/usr/local/lib/python3.8/dist-packages/nvmitten/debug/debug_manager.py", line 245, in _wrapper
    retval = obj(*args, **kwargs)
  File "/root/CM/repos/local/cache/db6b5d49a322435e/repo/closed/NVIDIA/code/retinanet/tensorrt/Retinanet.py", line 239, in create_network
    self.apply_subnetwork_io_types(network, subnetwork_name)
  File "/usr/local/lib/python3.8/dist-packages/nvmitten/debug/debug_manager.py", line 258, in _wrapper
    raise exc_info[1]
  File "/usr/local/lib/python3.8/dist-packages/nvmitten/debug/debug_manager.py", line 245, in _wrapper
    retval = obj(*args, **kwargs)
  File "/root/CM/repos/local/cache/db6b5d49a322435e/repo/closed/NVIDIA/code/retinanet/tensorrt/Retinanet.py", line 289, in apply_subnetwork_io_types
    self._set_tensor_format(tensor_in, use_dla=self.dla_enabled)
  File "/root/CM/repos/local/cache/db6b5d49a322435e/repo/closed/NVIDIA/code/retinanet/tensorrt/Retinanet.py", line 356, in _set_tensor_format
    tensor.allowed_formats = 1 << int(tensor_format)
TypeError: int() argument must be a string, a bytes-like object or a number, not 'NoneType'

Reproduce the calibration command for retinanet and start at closed/NVIDIA/code/retinanet/tensorrt/Retinanet.py, especially apply_subnetwork_io_types and _set_tensor_format, where the traceback shows a None tensor format. Determine why calibration reaches that value and verify that the command completes successfully after the issue is addressed.