OPERATORS__', '-DCMAKE_LIBRARY_OUTPUT_DIRECTORY=/mnt/home/ivan/second.pytorch/second/spconv/build/lib.linux-x86_64-3.6/spconv', '-DCMAKE_BUILD_TYPE=Release']-- Caffe2: CUDA
,我使用下面的安装caffe2,使用这个源中的命令,但是在训练模型时仍然显示了下面的错误。
我只需要知道在windows命令提示符上启用build_caffe2=1的命令。: Caffe2 support is not fully enabled in this PyTorch build.Please enable Caffe2 by building PyTorch from source with `BUILD_CAFFE2=1` flag.\Local\Programs\Python\Python310\lib\site-packa
错误报告包括:
In file included from /home/wangqm/anaconda3/conda-bld/caffe2-cuda9.0-cudnn7_1530109091069/work/caffe2for target 'caffe2/CMakeFiles/caffe2_gpu.dir/operators/caffe2_gpu_generated_depthwise_3x3_conv_op.cu.o' fai