Sergey Sandler

@yasakova-anastasia, this issue is a duplicate of https://github.com/opencv/cvat/issues/4707. The _Possible Solution_ described by the submitter of the issue #4707 works, just drop _sudo_. Inserting the following two lines at line...

@saikiran321, is the above link correct? I am getting the page not found error.

@saikiran321, _torch.onnx.export_ does not fail with OnnxExporterError: Unsupported with _saikiran321/pytorch_opset_20_. Unfortunately it now fails with _ValueError: only one element tensors can be converted to Python scalars_ after producing _UserWarning: ONNX...

@saikiran321, just curious about _saikiran321/pytorch_opset_20_ image. While _nvidia-smi_ reports _CUDA Version: 12.4_, `python3 -c "import torch; print(torch.cuda.is_available())"` prints _False_. Is this intentional?

@lkeab, the attached [Dockerfile](https://github.com/SysCV/MaskFreeVIS/files/11862653/Dockerfile.txt) helped to eliminate the problem. There are two minor issues with demo_video/demo.py, 1. modify line 162 to add a missing parameter, `predictions, visualized_output = demo.run_on_video(vid_frames, args.confidence_threshold)`...

@lkeab, with a modified [Dockerfile](https://github.com/SysCV/MaskFreeVIS/files/11953414/Dockerfile.txt) that includes CUDA 11.3 (not 11.1 as previously), and running CUDA kernels synchronously having `CUDA_LAUNCH_BLOCKING=1`, _demo_video/demo.py_ triggers either ` File "/MaskFreeVIS/demo_video/../mask2former/modeling/backbone/swin.py", line 159, in forward...

It seems to be possible to overcome the error reported above by modifying `BlockV2.__init__` by adding an _else_ clause after ``` if self.use_projection: self.proj_conv = nn.Conv2d( in_channels=channels_in, out_channels=channels_out, kernel_size=1, stride=stride,...

@sgjheywa, _scripting_ (`torch.jit.script`) helps to save a model with *dynamic* dimensions, while only _static_ dimensions are supported through _tracing_. There were many code changes to achieve JIT compatibility, please review...

@sgjheywa, the use case is _LibTorch_ integration in C++. The model can be compiled with `torch.compile`, but it does not help since you cannot save it with `torch.jit.save`. Am I...

@Cyril9227, _torch.onnx.export()_ fails for me too. It seems like the cause is described in https://github.com/pytorch/pytorch/issues/100790 that will be addressed through https://github.com/pytorch/pytorch/issues/114801 (ONNX opset 20 support). In the meantime I was...