[GitHub] [incubator-mxnet] junrushao1994 commented on issue #14244: Feature request: share memory between numpy.array and mxnet.ndarray

junrushao1994 commented on issue #14244: Feature request: share memory between numpy.array and mxnet.ndarray URL: https://github.com/apache/incubator-mxnet/issues/14244#issuecomment-481069471 @reminisce Sounds good. Will do!

[GitHub] [incubator-mxnet] junrushao1994 commented on issue #14244: Feature request: share memory between numpy.array and mxnet.ndarray

junrushao1994 commented on issue #14244: Feature request: share memory between numpy.array and mxnet.ndarray URL: https://github.com/apache/incubator-mxnet/issues/14244#issuecomment-481012596 Being lazy for a while, and now I am thinking of adding an API to mxnet. @wkcn @szha @SunDoge Wha

[GitHub] [incubator-mxnet] junrushao1994 commented on issue #14244: Feature request: share memory between numpy.array and mxnet.ndarray

junrushao1994 commented on issue #14244: Feature request: share memory between numpy.array and mxnet.ndarray URL: https://github.com/apache/incubator-mxnet/issues/14244#issuecomment-477808938 @wkcn this is because dlpack forgots to change its version number to 020 when releasing v0.2 lol.

[GitHub] [incubator-mxnet] junrushao1994 commented on issue #14244: Feature request: share memory between numpy.array and mxnet.ndarray

junrushao1994 commented on issue #14244: Feature request: share memory between numpy.array and mxnet.ndarray URL: https://github.com/apache/incubator-mxnet/issues/14244#issuecomment-477806143 @wkcn Sorry I made a mistake. We already got dlpack 0.2 in MXNet, so everything should be fine -

[GitHub] [incubator-mxnet] junrushao1994 commented on issue #14244: Feature request: share memory between numpy.array and mxnet.ndarray

junrushao1994 commented on issue #14244: Feature request: share memory between numpy.array and mxnet.ndarray URL: https://github.com/apache/incubator-mxnet/issues/14244#issuecomment-477781420 @wkcn I opened a PR about this to https://github.com/dmlc/dlpack/pull/38. --

[GitHub] [incubator-mxnet] junrushao1994 commented on issue #14244: Feature request: share memory between numpy.array and mxnet.ndarray

junrushao1994 commented on issue #14244: Feature request: share memory between numpy.array and mxnet.ndarray URL: https://github.com/apache/incubator-mxnet/issues/14244#issuecomment-477315300 I prototyped a version that supports zero copy from numpy to DLManagedTensor in dlpack, but it tu

[GitHub] [incubator-mxnet] junrushao1994 commented on issue #14244: Feature request: share memory between numpy.array and mxnet.ndarray

junrushao1994 commented on issue #14244: Feature request: share memory between numpy.array and mxnet.ndarray URL: https://github.com/apache/incubator-mxnet/issues/14244#issuecomment-471202979 Agreed with @wkcn This is an aut

[GitHub] [incubator-mxnet] junrushao1994 commented on issue #14244: Feature request: share memory between numpy.array and mxnet.ndarray

junrushao1994 commented on issue #14244: Feature request: share memory between numpy.array and mxnet.ndarray URL: https://github.com/apache/incubator-mxnet/issues/14244#issuecomment-469886372 The point is that we need transfer/share ownership of numpy's ndarray to mxnet's C++ backend, bec

[GitHub] [incubator-mxnet] junrushao1994 commented on issue #14244: Feature request: share memory between numpy.array and mxnet.ndarray

junrushao1994 commented on issue #14244: Feature request: share memory between numpy.array and mxnet.ndarray URL: https://github.com/apache/incubator-mxnet/issues/14244#issuecomment-469567431 Hey, I did some research this weekend and am convinced it might be possible. So is it acceptable