hudfdfdf

prroi_pooling_gpu.c seems not support to pytorch1.11 and pytorch1.12.

2

Hi prroi_pooling_gpu.c seems not support to pytorch1.11 and pytorch1.12. It has have the following questions： RuntimeError: Error building extension '_prroi_pooling': [1/3] /usr/local/cuda/bin/nvcc -DTORCH_EXTENSION_NAME=_prroi_pooling -DTORCH_API_INCLUDE_EXTENSION_H -DPYBIND11_COMPILER_TYPE=\"_gcc\" -DPYBIND11_STDLIB=\"_libstdcpp\" -DPYBIND11_BUILD_ABI=\"_cxxabi1011\" -isystem /usr/local/lib/python3.7/dist-packages/torch/include -isystem...

TypeError: `__cuda_array_interface__` must be a dict

6

File "/opt/research2/d3s/pytracking/tracker/segm/segm.py", line 96, in initialize feat_max_stride = max(self.params.features_filter.stride()) File "/opt/research2/d3s/pytracking/features/extractor.py", line 66, in stride return torch.Tensor(TensorList([f.stride() for f in self.features if self._return_feature(f)]).unroll()) TypeError: `__cuda_array_interface__` must be a dict I...

您好，首先感谢您的代码公开分享。我正在学习复现您的DFormerv2，源码训练NYUDepthv2数据集没有问题，但是在源码训练SUNRGBD数据集的时候，没有能够复现论文中的mIoU。和您源码的区别是，我使用了1张80G显存的A100显卡训练。附件是训练DFormerv2_S训练的LOG，希望可以指导一下，训练方式是否有问题，谢谢！ GPUS=1 NNODES=1 NODE_RANK=${NODE_RANK:-0} PORT=${PORT:-29122} MASTER_ADDR=${MASTER_ADDR:-"127.0.0.1"} export CUDA_VISIBLE_DEVICES="0" export TORCHDYNAMO_VERBOSE=1 PYTHONPATH="$(dirname $0)/..":"$(dirname $0)":$PYTHONPATH \ torchrun \ --nnodes=$NNODES \ --node_rank=$NODE_RANK \ --master_addr=$MASTER_ADDR \ --nproc_per_node=$GPUS \ --master_port=$PORT \ utils/train.py \ --config=local_configs.SUNRGBD.DFormerv2_S --gpus=$GPUS...