guanxinq commented on a change in pull request #17569: Adding sparse support to
MXTensor for custom operators
URL: https://github.com/apache/incubator-mxnet/pull/17569#discussion_r387352445
##########
File path: include/mxnet/lib_api.h
##########
@@ -1111,18 +1171,63 @@ extern "C" {
// create a vector of tensors for inputs
std::vector<MXTensor> inputs(num_in);
+
+ MXStorageType type;
+ void *data = nullptr;
+ void *data2 = nullptr;
+ MXSparse sparse;
+ MXSparse sparse2;
for (int i = 0; i < num_in; i++) {
- inputs[i].setTensor(indata[i], (MXDType)intypes[i], inshapes[i],
indims[i],
- inIDs[i], {indev_type[i], indev_id[i]});
+ // Dense representation.
+ if(!in_indices_shapes) {
+ type = kDefaultStorage;
+ data = indata[i];
+ }
+ // Sparse representation.
+ else {
+ // To do: remove if else.
+ if(!in_indptr_shapes) {
+ type = kRowSparseStorage;
+ sparse.set(indata[i], inshapes[i], indims[i], in_indices[i],
in_indices_shapes[i]);
+ }
+ else {
+ type = kCSRStorage;
+ sparse.set(indata[i], inshapes[i], indims[i], in_indices[i],
+ in_indices_shapes[i], in_indptr[i], in_indptr_shapes[i]);
+ }
+ data = (void*)(&sparse);
Review comment:
Thank Sam for the comments.
1. The complicated initialization sequence is simplified in the latest
update.
2. If we manage the MXSparse in MXTensor, we need to add extra MXSparse
variables to MXTensor, which is similar to my current fix.
3. I know the overwrite issue which is fixed now.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
